[CHROMOSOME ORGANIZATION AND GENIC EXPRESSION |
BARBARA|McCLINTOCK

Department of Genetics, Carnegie Institution of Washington, Cold Spring Harbor, N. Y.

During the past six years, a study of the be-
havior of a number of newly arisen mutable loci
in maize has been undertaken. This study has
provided a unique opportunity to examine the mu-
tation process at a number of different loci in
the chromosomes. For some of these loci, sev-
eral independent inceptions of instability have
occurred during the progress of this study. The
types of mutation that appear, and the types of
instability expression, need not be the same at
any one locus. In fact, comparisons of the be-
havior of these different mutable conditions at a
particular locus have shown. striking ‘diversity,
not only with regard to the changes in phenotypic
expression that result from mutations at the locus,
but also with regard to the manner in which muta-
bility is controlled. Knowledge of the genetic
constitutions, with respect to mutable loci al-
ready present in the plants in which new mutable
loci have arisen, and the subsequent behavior
of the newly arisen mutable loci, have provided
evidence that allows an interpretation of their
mode of origin and also their mode of operation.
As a consequence of this study, some rather un-
orthodox conclusions have been drawn regarding
the mechanisms responsible for mutations arising
at these loci. The same mechanisms may well
be responsible for the origins of many of the ob-
served mutations in plants and animals.

Instability of various loci—whether referred to
by the terms mutable loci, mutable genes, or
variegation, position effect, etc.—has been known
for many years, and many such cases have re-
ceived considerable study. The conditions as-
sociated with the more obvious position-effect
phenomena in Drosophila are well known. Those
associated with instability of phenotypic ex-
pressions in other organisms have been less well
understood. It is because of the distinctive ad-
vantages that the maize plant offers for such a
study that it has been possible to obtain precise
evidence concerning some of the events associated
with the origin and behavior of mutable loci. The
first of these advantages relates to the ease of
observing the chromosomes, and thus determining
the nature of some of the changes that occur in
them. The presence of a triploid endosperm in

the kernel provides a second advantage. This
endosperm, with its outer aleurone layer that can
develop pigments, and the underlying tissues that
may develop starches of several types, or sugars,
or carotenoid pigments, permits the detection of
differences in phenotypic expression of various
types. Some of these may be quantitatively meas-
ured, Thirdly, there are a number of different loci
known in which heritable alterations have given
rise to changes in the expression of these several
endosperm components. The mutations at some of
these loci affect characters of both the endosperm
and the plant tissues. This applies particularly
to those mutations that affect the development
of the anthocyanin pigments. In the studies to
be described, the presence in the short arm of
chromosome 9 of four marked loci that affect
endosperm characters has been of particular im-
portance for analyzing the events occurring at
mutable loci. The necessity of having such mark-
ers will become evident in the discussion. For
this study, the accumulated knowledge of the be-
havior of newly broken ends of chromosomes in
maize has been of particular importance. Its
significance for interpreting the origin of mutable
loci will be indicated in the sections that follow.

THE CHROMATID AND CHROMOSOME TYPES OF
BREAKAGE-FUSION-BRIDGE CYCLE

The diagrams of Figure 1 illustrate the mode
of origin of newly broken ends of chromosomes
at a meiotic mitosis and the subsequent behavior
of these ends in successive mitotic cycles. A
chromosome with a newly broken end entering a
telophase nucleus in the gametophytic or endo-
sperm tissues will give rise in the next anaphase
to a chromatid bridge configuration (McClintock,
1941). The bridge is produced because fusion
occurs between sister chromatids at the position
of previous anaphase breakage. This sequence
of anaphase breaks and sister-chromatid fusions
will continue in successive mitoses. It has
therefore been designated the chromatid type of
breakage-fusion-bridge cycle. This cycle is il-
lustrated in A of Figure 1. In the sporophytic
tissues, however, this cycle usually does not
occur. The broken end entering a telophase

[13]

Arman ttn Rae

 
14 BARBARA McCLINTOCK

THE GHROMATIO TYPE OF BREAKAGE -FUSION- BRIDGE CYCLE
MEIOSIS: WICROWPORE OR WEGASPORE
Prophose synapsis (one type} Prophase
1 —_——_— 4 —=—

Graspover (ene type)
a

First meiotic onaphase

Anaphare

: wna (

Following prophase
6.

 

THE CHROMOSOME TYPE OF BREAKAGE -FUSION-SRIDGE CYCLE: SAME AS A FROM
MEIOSIS TO FORMATION OF EGG AND SPERM NUCLEI

Zygote nucleus Fusion of broken ends in each telophese nucious

—S-— Cc TS ‘0

Firs! onaphese

od

Telophase of firet division in embryo

on™, AN 2 cro ) wa
. : NY
. a or
V7 NY

FIG. 1. Diagrams illustrating the origin of a newly
broken end of a chromosome at the meiotic anaphase
and its subsequent behavior. A. The chromatid type
of breakage-fusion-bridge cycle. 1. One type of syn-
aptic configuration at the first meiotic prophase be-
tween homologous arms of a pair of chromosomes, one
member of which carries a duplication of this arm in
the inverted order. 2. The production of a dicentric
chromatid as the consequence of a crossover. It is
composed of two complete chromatids of this chromo-
some. 3. Anaphase /. Bridge configuration produced
by separation of centromeres of the dicentric chromatid.
A break in the bridge occurs at some position between
the two centromeres. 4. Fusion of sister chromatids
at the position of the previous anaphase break is ex-
hibited in prophase of the microspore or megaspore
nucleus. 5. Separation of sister centromeres at ana-
phase in the microspore or megaspore produces a
bridge configuration. This bridge is broken at some
position between the two centromeres. 6. Fusion of
sister chromatids occurs at the position of the pre-
ceding anaphase break. Separation of sister centro-
meres at the next anaphase again produces a bridge
which is broken at some position between the two
centromeres. This cycle continues in successive
mitoses during the development of the gametophyte
and the endosperm.

B. The chromosome type of breakage-fusion-bridge
cycle. It may be initiated in the sporophyte if each
gamete contributes a chromosome which has been
broken in the anaphase of the division preceding gamete
formation. The zygote nucleus will then contain two
such chromosomes. In the prophase of the first division
of the zygote, 7, each of these is composed of two
sister chromatids fused at the position of the previous
anaphase break. In the first anaphase of the zygotic

——— oo

Following prophase

Soo
1 ————

Anaphase

nucleus heals, and its subsequent behavior re-
sembles that of a normal, nonbroken end of a
chromosome. (Note: The chromatid type of
breakage-fusion-bridge cycle can continue through-
out the development of the sporophytic tissues
under certain conditions. These conditions are
usually not present in the genetic stocks of
maize.) If, however, a chromosome with a newly
broken end is introduced into the zygote by each
gamete nucleus, the broken ends of the two chro-
mosomes are capable of fusion (McClintock, 1942).
This establishes a dicentric chromosome. A dif-
ferent type of breakage-fusion-bridge cycle is
thereby initiated. In the telophase nuclei, the
fusions now occur between the broken ends of
chromosomes rather than between the broken ends
of sister chromatids, as described above. This
sequence of events has been called the chromo-
some type of breakage-fusion-bridge cycle, and
is illustrated in B of Figure 1. A study of the
consequences of these cycles has revealed that
they may initiate breakage events in chromosomes
of the complement other than those undergoing
the cycle. This complication has been of signif-
icance, for it appears that these unanticipated
alterations of the chromosomes may be respon-
sible primarily for the origin of mutable loci and
of other types of heritable change.

UNEXPECTED CHROMOSOMAL ABERRATIONS
INDUCED BY THE BREAK AGE-FUSION-
Bripcre CyYcLes

In the course of an experiment designed to in-
duce small internal deficiencies within the short
arm of chromosome 9, a number of plants were
obtained that had undergone the chromosome type
of breakage-fusion-bridge cycle in their early
developmental period. The short arm of each
chromosome 9 was involved in this cycle. It is

 

division, 8, these two chromosomes give rise to bridge
configurations as the centromeres of the sister chro-
matids pass to opposite poles. Breaks occur in each
bridge at some position between the centromeres. In
the telophase nuclei, two chromosomes, each with a
newly broken end, are present as diagrammed in 9.
The crosses mark the broken ends of each chromo-
some. Fusion of broken ends of chromosomes occurs
in each telophase nucleus, 10, establishing a dicen-
tric chromosome. In the next prophase, 11, each
sister chromatid is dicentric. At the subsequent ana-
phase, several types of configurations may result
from separation of the sister centromeres, two of
which are shown in 12, Separations as shown in 6
of 12 give rise to anaphase bridge configurations.
Breaks occur in each bridge at some position between
the centromeres. The subsequent behavior of the
broken ends, from telophase to telophase, is the same
as that given in diagrams 9 to 12.
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 715

known that the cycle will often cease suddenly
in certain cells and that these cells are then
capable of developing sexually functional branches
of the plant. In order to determine the nature of
the chromosome changes produced by this cycle,
the sporocytes of many of these plants were ex-
amined at the pachytene of meiosis. The expected
types of altered constitution of the short arm of
chromosomes 9 were found. In addition, other
quite unexpected types of chromosome aberration
appeared in a number of the plants. These al-
terations had been produced in the early develop-
mental periods when the breakage-fusion-bridge
cycles were occurring. With a few exceptions,
the chromosome. parts in which alterations had
been initiated were the knobs and the centromeres,
or the nucleolus organizer of chromosome 6. ° In
the majority of cases, either the knob or the centro-
mere of one of the chromosomes 9 that had been
undergoing the breakage-fusion-bridge cycle was
involved in the structural rearrangement. Non-
randomness was apparent with regard to the other
chromosome involved in the aberration.
ample, four cases were found in which the centro-
mere of chromosome 9 had fused with the centro-
mere of another chromosome—chromosome 2 in
three of the four cases. Chromosome 8 was also
very frequently involved in these structural
changes.

The breakage-fusion-bridge cycle was obviously
responsible for the induction of these alterations

in the knobs, centromeres and the nucleolus or- |

ganizer. That alterations in such elements were
occurring without obvious direct participation
of the knob or the centromere of the chromosome
9 undergoing the breakage-fusion-bridge cycle
has also been indicated. This was made evident
by the presence in one plant of an inversion in-
volving the nucleolus organizer and the centro-
mere region in chromosome 6, by an inversion in
chromosome 5 in another plant involving the
centromere and the knob regions, and by an in-
version in chromosome 7 in a third plant involv-
ing the centromere region and the knob region in
the long arm of this chromosome. In addition,
some of the plants examined showed the presence
of a ring chromosome that was not composed of
segments of chromosome 9, so far as could be
determined. It now must be emphasized that it
was in the self-pollinated progeny of plants that
had undergone the chromosome type of breakage-
fusion-bridge cycle in their early developmental
Period that the initial burst of newly arisen mu-
table loci appeared. It might be suspected that
this burst was a reflection of the mechanism that

For ex- |

had produced the alterations mentioned above.
If so, the origin of mutable loci would be asso-
ciated with change in these particular elements
of the chromosome complement. It was some time,
however, before sufficient evidence had accu-
mulated to allow deductions to be drawn regarding
this presumptive relationship. A description of
the origin and behavior of some of the representa-

_ tive types of mutable loci should be given before

this topic is again considered.

RECOGNITION OF THE RELATION OF MUTATION
, TO THE MiToTIc CYCLE

Interest in these mutable loci,, appearing un-

. expectedly and in large numbers in the self-

pollinated progeny of plants that had undergone
the chromosome type of breakage-fusion-bridge
cycle in their early developmental periods, was
aroused when it was realized that in each case
some factor was present which controlled the
time or the frequency of mutations. This factor
could be altered as a consequence of some event
associated with the mitotic process. This was
made evident by the appearance of sectors of
tissue, derived from sister cells, that exhibited
obvious differences in time of mutations, mutation
frequency, or both. In many cases, it was also

. apparent that the mutations themselves arose as

a consequence of some event associated with the
mitotic cycle. This basic behavior pattern was
exhibited by all the various newly arisen mutable
loci. It directed attention to the mitotic mecha-
nism as the responsible agent. It was concluded,
therefore, that further investigation of these mu-
table loci might produce some evidence leading
to an appreciation of the nature of the responsible
mitotic events.

During six years of study of a number of newly
arisen mutable loci, some well-established facts
have accumulated concerning the processes asso-
ciated with the origin of mutable loci and their
subsequent behavior. Observation of consistent
behavior in many mutable loci, where the cyto-
logical events associated with a change in pheno-
type could be determined, and comparison of the
behavior of these loci with others in which cyto-
logical determinations could not readily be made,
have provided an assemblage of interrelated facts
upon which the conclusions to be stated later
are based.

THE OrIGIN OF Ds AND ITS BEHAVIOR

The first evidence of the type of chromosomal
event that is associated with the expression of
mutability came with the discovery of a locus in
16 BARBARA McCLINTOCK

the short arm of chromosome 9 at which chromo-
some breaks were occurring. This was observed
in the self-pollinated progeny of one of the plants
that had undergone the chromosome type of
breakage-fusion-bridge cycle in early development.
When first seen, the ‘“‘mutability’’ was expressed
by the time and frequency of the breaks that oc-
curred at this locus in some cells during the
development of a tissue. Also, some change
could occur in somatic cells that affected the
time and frequency; and this latter event like-
wise was associated with the mitotic process.
The behavior pattern resembled in considerable
detail the patterns exhibited by the mutable loci.
In this case, however, a mechanism associated
with chromosome fusion and subsequent breakage
was responsible for the behavior observed. The
mutations from recessive to dominant exhibited
by the mutable loci would not alone have sug-
gested a chromosome-breakage mechanism as
being responsible. Because of this similarity
of the patterns of behavior, it was suspected that
the basic mechanism responsible for mutations
at mutable loci could be one associated with some
form of structural alteration at the locus showing
the mutation phenomenon. This conclusion was
consistent with the very first observations of the
behavior of mutable loci. These observations
had indicated that the events at mutable loci
leading to mutations and also other events con-
“trolling their time and frequency of occurrence
were associated with alterations that were in
some manner produced during the course of a
mitotic cycle.

Intensive study of this locus in chromosome
9 at which structural alterations occur at regulated
rates and at regulated times in development has
been rewarding. A “‘break’’ in the chromosome
at this locus was the event first recognized. The
factor responsible was therefore given the symbol
Ds, for “‘Dissociation.”” The nature of the break-
age event was later determined. It arises from
dicentric and acentric chromatid formations. The
acentric fragment is composed of the two sister
chromatids, from the Ds locus to the end of the
short arm.. The complementary dicentric com-
ponent includes the sister segments from the
locus to the centromere plus the long arms of
the two sister chromatids. This is the type of rec
ognizable event found to occur most frequently at
Ds. Other recognizable aberrations, however,
may sometimes arise. One of them is the forma-
tion of an internal deficiency in the short arm
of chromosome 9. Such deficiencies include the

regions adjacent to Ds, and vary in extent from
minute to quite large. Translocations between
this chromosome and another chromosome of the
complement may arise, with one of the points of
breakage at the Ds locus. Duplications, or in-
versions, of segments within chromosome 9 may
also be produced, one of the breakage points
being at Ds.

It was realized early in this study of Ds that
changes could occur at the locus leading to
marked alterations in frequency of the detectable
breakage events. The original isolate was show-
ing high frequencies of formation of dicentric
chromatids and the associated acentric fragments.
Changes arose at the locus, however, aS a con-
sequence of some event occurring in a somatic
cell. These changes resulted in the appearance,
in subsequent cell and plant generations, of
lowered frequencies of these events. Such changes
in the behavior pattern of Ds were called ‘*changes
in state’’; and the Ds with the altered state be-
haved in inheritance as an allele of the original
isolate of Ds. A subsequent change could occur,
which again was recognized by an altered fre-
quency of detectable breakage events, and which
behaved in inheritance as an allele of the initial
state, of the derived state, or of other unrelated
derived states. By selecting altered states of
Ds, a series of alleles of the original Ds has
been isolated. The changes in state of Ds, and
those occurring at other mutable loci, are of con-
siderable significance in understanding the nature
of the events responsible for the patterns of be-
havior of all mutable loci. A discussion of this
significance will be postponed until the behavior
of some other mutable loci have been considered.

The meaning of the term will then be readily
apparent.

TRANSPOSITION OF Ds

An important aspect of this study, with regard
to the origin of mutable loci and nature of their
mutation process, is related to transposition of
Ds from one location in the chromosome comple-
ment to another. The discovery of such trans-
positions occurred in the course of studies aimed
at determining the exact location of Ds in chro-
mosome 9. These tests involved linkage relation-
ships. A sequence of six marked loci along the
chromosome arm were used, and the linkage
studies clearly established the location of Ds
as shown in Figure 2. This genetically deter-
mined location fitted the position of breaks in
the chromosome observed in some of the sporo-
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 17

cytes of plants having Ds in either one or both
chromosomes 9. Such chromosome breaks are
illustrated in the photographs of microsporocytes
at pachytene given in Figures 4 to 8. This was
the location of Ds when it was first discovered,
and has been called the standard location.

In the course of studies of the inheritance be-
havior of Ds, an occasional kernel appeared
which showed that Ds-type activity—that is,
chromosome breakage—was occurring at a new
position in the short arm of chromosome 9. At-
tempts were made to germinate such kernels when
they were found. If a plant arose from one, a
study was then commenced to determine the new
location of the Ds-type activity. Over 20 cases
of the sudden appearance of Ds-type activity in
new locations in the short arm of chromosome 9,
and several cases of its sudden appearance in
other chromosomes of the complement, have been
investigated. Within the short arm of chromosome
9, such activity has appeared at various positions.
All the isolates studied have shown sharply de-
fined locations of the Ds-type activity. In these
cases, the cytological determination of. breakage
position and the genetic determination of location
were in agreement. New positions of Ds-type
activity have appeared between all of the marked
loci shown in Figure 2. For example, in four in-
dependently arisen cases, the new position of
Ds has been located between / and Sh, In two of
these, it is to the right of /, at or close to the
same position in each case—approximately one-
fifth the crossover distance between / and Sh.
In the other two it is to the left of Sh, with a
very low percentage of crossing over between Ds
and Sh in each case.

The mode of detecting new locations of Ds-
type activity has been selective, in that those
arising in the short arm of chromosome 9 are im-
mediately revealed on many of the ears coming
from test crosses. Ds-type activity has suddenly
appeared, however, in other chromosomes of the
complement. Only when appropriate genetic
markers are present can it be detected readily;
and in most tests, such markers have not been
present.

Several questions must now be asked, How do
new positions of Ds activity arise? And what
conditions are responsible for their occurrence?
The methods used in seeking answers to these
questions may be described. In some cases, it
averted established that the appearance of Ds
ite ed at a new location was associated with
Ithee appearance at the known former location.

emphasized that the mechanism under-

1

knob Yq G Sh Bz

 

Wx Os centromere
L/ Lid Ll L
o— + + ———
¥q ¢ sh bz wr

FIG. 2. Diagram showing the approximate locations
of the genetic markers in the short arm of chromosome
9 that have been used in this study. In symbolization,
dominance is indicated by a capital letter or capital-
ization of the first letter. Recessiveness is indicated
by lower-case letters, The symbols refer to the fol-
lowing plant or endosperm characters: Yg, normal
chlorophyll; yg, yellow-green chlorophyll color in
early period of development of the plant. Sh, normal
endosperm; sh, shrunken endosperm. I, C, and ¢ form
an allelic series associated with pigment development
in the aleurone layer of the endosperm. J, inhibitor of
aleurone color formation, dominant to C. C, aleurone
color, dominant to c, colorless aleurone. The Bz
factor is associated with development of aleurone and
plant color. When homozygous, the recessive, bz,
(bronze), gives rise to an altered anthocyanin color in
the aleurone and plant tissues, from a dark red or
purple to a bronze shade. When Wx is present, the
starch in the pollen and endosperm stains blue with
iodine solutions, due to the presence of amylose
starch; when only the recessive wx (waxy) is present,
no amylose starch is formed and with iodine solutions,
the starch stains a reddish-brown color. The position
of Ds, indicated in the diagram, is the standard loca-
tion (see text).

“lying Ds events is one that can give rise to’ trans-
locations, deficiencies, inversions, ring-chromo-
somes, etc., as well as the more frequently oc-
curring dicentric chromatid formations with
reciprocal formation of acentric fragments. It
has also been stated that in each such case one
breakage point is at the known location of Ds.
The appearance of Ds at new locations is prob-
ably associated with such a break-inducing mech-
anism. This was indicated by extensive analysis
of the constitutions of two independent duplica-
tions of segments of the short arm of chromosome
9 when a new location of Ds activity was also
present in this arm. In both cases, only one of
the many tested gametes of one of the parent
plants carried the particular chromosome aber-
ration with the new location of Ds. It was de-
tected in two single aberrant kernels on separate
ears coming from similar types of crosses made
in two different years. The female parent carried
two morphologically normal chromosomes 9, each
with the markers C, sh, bz, and wx. No Ds (or
Ac, see below) was present in these plants. The
male parent (one Ac present) carried two morpho-
logically normal chromosomes 9. The markers
I, Sh, Bz, Wx, and Ds (at its standard location)
were present in one chromosome 9. The homol-
ogous chromosome carried C, sh, bz,. wx, but no
18

<—a

BARBARA McCLINTOCK

 

c |
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION

 

FIG. 3, Photograph of a normal bivalent chromosome 9 at pachytene of meiosis.
diagram, 3a, the knob terminating the short arm is indicated by the arrow, a, The centromere is indicated
by the arrow 6, Mag. approximately 1800x. Fusion of homologous centromeres appears to occur at pachy-
tene. Consequently, in the diagrams accompanying Figures 3 to 8, this region is indicated as single rather
than double. :

FIGS. 4 to 7 and accompanying diagrams, 4a to 7a. I!lustrations of the position of breaks at the Ds locus
as seen at pachytene of meiosis in plants having Ds at its standard location in one chromosome 9 and no
Ds in the homologue. The two homologues are distinguishable. At the end of the short arm of the chromo-
some 9 having no Ds, a segment of deep-staining chromatin extends beyond the knob. The short arm of the
chromosome 9 carrying Ds terminates in a knob. Magnifications approximately 1800x. In Figure 4, a break
at Ds occurred in a premeiotic mitosis. The acentric fragment, from Ds to the end of the arm, was lost to
the nucleus, Consequently, this segment is missing in the bivalent. The homologous segment in the
chromosome 9 having no Ds is therefore univalent. In making the preparation, this segment was considerably
stretched. In the accompanying diagram, arrow a points to the knob and the small deep staining segment
extending beyond the knob. Arrow 6 points to the centromere region, not clearly shown in the photograph.
Arrow ¢ points to the position of the break in the chromosome 9 that carried Ds, Figures 5 and 6 show the
appearance of the bivalent chromosome 9 when a break in the Ds carrying chromosome occurred at the
meiotic prophase and when the free segment, from Ds to the end of the arm, paired with its homologous
Segment in the chromosome 9 having no Ds, In the accompanying diagrams, arrow a points to the knobs,
arrow 6 points to the centromeres and arrow ¢ to the position of the Ds break in one of the homologues.
Figure 7 is similar to Figures 5 and 6 except that the free fragment, from the position of Ds to the end of
the arm, did not pair with its homologous segment in the chromosome 9 having no Ds. In the accompanying
diagram, arrow a! points to the knob and the deep-staining chromatin extending beyond the knob in the chro-
mosome having ne Ds. Arrow a? points to the knob of the unpaired acentric fragment. Arrow b points to
the position of the centromeres, not observable in the photograph. Arrow c? points to the broken end of the
centric segment, and arrow c? points to the broken end of the acentric segment. (For Fig. 7, see next page.)

In the accompanying

19
20 BARBARA McCLINTOCK

 

7
“
al 7a
c!l—>
<b
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 21

Whe” “N .
¥ ns

™~

  

8a

—b

f

c

FIG. 8, The chromosome 9 bivalent at pachytene in a plant having Ds at its standard location
in each chromosome 9, A Ds break occurred at the meiotic prophase in each chromosome. Con-
sequently, the bivalent is composed of two free segments: one short acentric segment, from Ds
to the end of the arm, and a long centric segment, from the position of Ds to the end of the long
arm. In the accompanying diagram, arrow a points to the large knob terminating the short arm of
each member of the acentric bivalent segment, and arrow 6 points to the centromeres of the
centric bivalent segment. Arrow c' points to the broken ends of the centric bivalent segment
and arrow c* points to the broken ends of the acentric bivalent segment.

Ds. The constitutions of the aberrant chromosomes
9, which were present in an individual pollen
grain in each case, are shown in Figure 9. Each
has a duplication of a segment of the short arm
of chromosome 9, A study of these constitutions
reveals that, in each case, the new position of
Ds activity coincided with the position of one
of the breaks that produced the duplication. Also,
the position of the second break coincided with
the previously known location of Ds in the mor
Phologically normal Ds-carrying chromosome 9
of the male parent plant. It seems clear from the
analysis of both these cases that the breaks

were Ds initiated, and also that both breaks in-
volved sister chromatids at the same location.
Two other cases of newly arisen duplications
associated with new positions of Ds activity
have received study. These have proved to be
similar in their modes of origin to the two cases
diagrammed. Although these cases suggest that
the new positions of Ds activity may arise from
contacts with the Ds that is present in the chro-
mosome complement, they do not constitute evi-
dence that Ds is composed of a material sub-
stance and that the new positions arise from in-
sertion of this material into new locations.
22 BARBARA McCLINTOCK

 

1 Ds Sh & Wi We Bz Sh De
WZ UY _
Q
wt Ee

 

FIG. 9. Diagrams illustrating the constitution of
two chromosomes 9, each of which carries a duplicated
segment (heavy line). A transposition of Ds accom-
panied the formation of the duplication in each case,
For details, see text.

Ac BEHAVIOR AND INHERITANCE: GENERAL
CONSIDERATIONS

Before continuing discussion of the sudden ap-
pearance of Ds at new locations, it is necessary
to consider another heritable factor called Acti-
vator (Ac). Ac is a dominant factor that must be
present if any breakage events are to occur at
Ds. If Ac is absent from the nucleus, no detect
able events whatsoever occur at Ds, wherever it
may be located; and no new positions of Ds ac-
tivity appear. Because of this, and because new
positions of Ds arise only in a few cells, during
the period of development of a tissue when
breakage events are occurring at Ds, it may be
concluded that these new positions are one of the
consequences of the mechanism that produces
chromosome breakage events at Ds.

The location of Ds in chromosome 9 was
first suggested by the altered phenotype of
sectors of tissue derived from cells in which a
break had occurred at Ds. Such sectors will ap-
pear if the factorial constitutions of the homo-
logues differ. The chromosome arm carrying Ds
must also carry dominant markers, and the homo-

logue must carry the recessive alleles. If one
chromosome 9, delivered by the male parent, car-
ries I, Sh, Bz, Wx, and Ds at its standard loca-
tion, and if the homologue delivered by the fe-
male parent carries the recessive alleles C, sh,
bz, and wx, and has no Ds, the events at Ds that
form a dicentric chromatid and an acentric frag-
ment will lead to elimination of the acentric frag-
ment during a mitotic division. This fragment
will form a pycnotic body in the cytoplasm, which
subsequently disappears. All the dominant mark-
ers carried in this acentric fragment will be re-
moved from the nuclei after such an event has
occurred, The sector of tissue derived from these
cells will exhibit the collective phenotype of
the recessive alleles carried by the homologous
chromosome 9 having no Ds.

If pollen of plants having one Ac factor and
carrying 1, Sh, Bz, Wx, and Ds (standard location)
is placed upon silks of plants carrying C, sh, bz,
and wx in their chromosomes 9, but having no Ds
or Ac, two types of kerels will appear, those
that received an Ac factor and those without Ac.
The latter kernels will be colorless and non-
shrunken, and will show the Wx phenotype. No
variegation for the characters exhibited by the
recessive alleles will appear, as shown in Figure
10. If Ac is present in the endosperm, however,
the described Ds events will occur in some cells
during the development of the kernel. This will
result in elimination of the segment in the short
arm of chromosome 9 carrying the dominant
factors. All the cells arising from one in which
such an event has occurred will exhibit the C,
sh, bz, and wx phenotypes. Consequently, these
kernels will be variegated. The photographs of
kernels in Figures 11 to 13 will illustrate the
nature of this variegation.

FIGS. 10 to 15. Photographs of kemels illustrating the effects produced by breaks at Ds, and the relation
of the presence or absence of Ac and the doses of Ac on the time of occurrence of Ds breaks. The kemels
arose from the cross of a plant (9) carrying C and bz and having no Ds in each chromosome 9, by plants (¢)

having /, Bz, and Ds at its standard location in chromosome 9.
completely colorless due to the inhibition of aleurone color when | is present.

present.

In Figure 10, no Ac is present; the kernel is
In Figures 11 to 13, one Ac is

Breaks at Ds occur early in development with consequent loss of / and Bz to the nuclei. The cells
without J and Bz give rise to sectors showing the C bz phenotype.

It should be noted that Bz substance dif-

fuses through several cell layers, from the / Bz areas into the C bz areas, producing a C Bz phenotypic ex-
pression in these genotypically C 6z cells. Thus, all large areas of the C bz genotype are rimmed with the
dark aleurone color of the C Bz phenotype. Small C bz areas may be mostly C Bz in phenotype because of this
diffusion, and very small C bz areas are totally C Bz in phenotype. These relationships are clearly expressed
in the photographs. In Figures 11 to 13, both large and small C bz sectors are present as well as areas in
which no C bz sectors appear. This irregularity arises from alterations that occur to Ac during the development
of the kernels. These give rise to sectors with no Ac or with altered doses of Ac (see text). Figure 14 shows
the pattern that may be produced by Ds breaks when two Ac factors are introduced into the primary endosperm
nucleus, There are many small sectors of the C bz genotype, all produced by relatively late occurring Ds breaks.
This results in a heavily speckled variegation pattern, Figure 15 shows the pattern that may appear when
three Ac factors are present. Only relatively few C bz specks, produced by late occurring Ds breaks, are pres-

ent in this kernel.
23

CHROMOSOME ORGANIZATION AND GENIC EXPRESSION

 
24 BARBARA McCLINTOCK

In the early studies, it soon became apparent
that the time ofoccurrence of Ds breakage events
in the development of a tissue depends upon the
dose of Ac present. The higher the dose of Ac,
the later in development will such events at Ds
occur. Because the endosperm is triploid—the
female parent contributing two haploid nuclei
derived from the female gametophyte, and the
male contributing one haploid nucleus—single
to triple doses of Ac may readily be obtained
after appropriate crosses. The time of response
of Ds to double and triple doses of Ac is shown
in Figures 14 and 15.

Initial studies of the inheritance behavior of
Ae showed that it follows the mendelian laws
known to apply to single genetic units, An illus-
tration of this inheritance behavior is given in
Figure 16. The ear in this figure was derived
from a cross in which the female parent carried
the recessive factor c in each chromosome 9, but
had no Ds or Ac factors, and the male parent
carried in each chromosome 9 the factors C and
Ds (standard location). It also carried a single
Ac factor. The expected 1-to-l ratio of colored,
nonvariegated kernels (no Ac present) to kernels
showing sectors of the c phenotype (Ac present)
was apparent on this ear. Efforts to determine
the location of Ac in the chromosome complement
were commenced, but were soon abandoned when
it was realized that Ac need not remain at any
one location in the complement. It can appear
at new locations and in different chromosomes.
Because of this highly unexpected behavior of a
genetic factor, extensive studies were made to
determine the mode of inheritance of Ac and to
learn how these new positions arise. Much has
been learned from them about Ac behavior and
inheritance. The modes of investigating altera-
tions of Ac will be considered later. Some of
the facts concerning its behavior, however, may
be stated here in summary form, by describing
the results obtained from several types of
experiments.

If plants having one Ac factor are self-pollinated,
the expected mendelian ratios may appear in the
F, populations. These are: one with two Ac, to
two with one Ac, to one with no Ac. The ratios
obtained in one such test were 61: 145: 68, which
is close to the statistical expectancy. When,
however, these F, plants having two Ac factors
are crossed by plants having no Ac, the expecta-
tion would be that all the progeny would have
one Ac factor. Usually, most of the plants do
have one Ac factor; but sometimes there are
plants displaying other conditions with respect

 

FIG. 16. Photograph of an ear produced when pollen
from plants having C and Ds (standard location) in
each chromosome 9 and carrying one Ac factor is
placed on silks of plants carrying ¢ in each chromo
some 9 and having no Ds or Ac factors. Approximately
half of the male gametes have no de factor. The
kernels arising from the functioning of such gametes
are fully colored; no variegation appears. The other
half of the male gametes carry Ac. The kernels arising
from the functioning of these gametes are variegated.
They show a number of sectors with the ¢ phenotype.
Note the approximate 1 to 1 ratio of fully colored ker-
nels to variegated kernels.

to Ac. The following unexpected types have ap-
peared: plants having no Ac factor; others having
two nonlinked Ac factors; others having two Ac
factors that appear to be linked; still others
having an Ac factor that acts as a single unit in
inheritance but gives the same dosage action as
would two doses of the Ac factor in the parent
plant that contributed Ac. The dosage action of
Ac may be altered in other ways, moreover, so
that a single Ac factor, as determined genetically,
may exert an action either less than that of the
Ac factor contributed by the Ac-carrying parent
plant or falling between one and two doses of
this. factor.

In the early studies of Ac inheritance, the Ac
factor was found not to be linked with the genetic
markers in the short arm of chromosome 9. In
one series of tests of Ac inheritance, where a
number of Ac-carrying plants were derived from
a cross between a plant having no Ac and a plant
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 29

having one Ac, a single aberrant plant was pres-
ent in the F,. Tests of this plant showed that it
possessed a single Ac factor, which was obviously
linked to the markers in the chromosome 9 de-
livered by the Ac-carrying parent plant. In the
sister plants, and in the parent plant that con-
tributed Ac, no such linkage was evident. Studies
were then conducted to examine the linkage re-
lationships of this Ac with the marked loci. The
very same types of tests for linkage of Ac with
the genetic markers in chromosome 9 were also
conducted with the sister plants. A summary of

tance to the left of J, Studies were then under-
taken to investigate not only the linkage behavior
of Ac but also the types of events that alter Ac
in these two sharply delimited locations. It was
determined that in the majority of the Ac-carrying
gametes produced by plants having an Ac at
either of these two stated positions, no change
in location or action of Ac occurred. Exchanges
of Ac from one homologue to another took place
as a consequence of crossing over, with consistent
frequencies in each case. In a few of these
gametes, however, the above-described types of

TABLE 1

Comparisons of Ae inheritance: (1) when Ac was not linked to markers in chromosome 9,
and (II) when Ac was linked to markers in chromosome 9, in crosses of

() gi sa Bz Wx Ds 1 Ae,

 

 

 

a nonlinked
C Sh Bz wx Ds
9 C'sh bz wx/C sh bz wx, no Ds, no Ae by (1) and
(iD ol Sh Bz Wx Ds Ae
C Sh Bz wx Ds ac
I Il
Chromosome-9 Constitution Variegated* Nonvariegated* Variegated* Nonvarie gated *
of dGamete Ae present No Ae Ac present No de
I Wx 268 255 928 246
C wx 248 242 164 893
I wx 88 91 62 387
C Wx 84 83 344 100

 

*Presence of Ac detected by sectors of the C sh bz wx phenotype in kernels on ears obtained from cross I or
cross Il, In the absence of Ac, no Ds events occur; the kernels are therefore nonvariegated,

one set of these comparative tests of Ac inherit-
ance is presented in Table 1. In part I of this
table, no linkage of Ac with / or Wx is evident;
in part II, however, linkage is obvious. The data
place Ac to the right of Wx. Approximately 20 per
cent crossing-over appears to have occurred be-
tween Wx and Ac. Actually, this figure is only
approximate, for a few kernels in the Ac-wx class
do not carry Ac in chromosome 9. They do not
belong in the crossover class. The Ac in these
kernels had been transposed from chromosome 9
to another chromosome. Also, a few kernels in
the Wx-no Ac class have not lost Ac because of
crossing over but rather because it was removed
from its location in chromosome 9.

The described case of sudden appearance of
Ac in chromosome 9 is not the only one that has
been found and similarly studied. Seven independ-
ent cases have so far been identified. In two of
these, Ac appeared in the short arm of this chro-
mosome: in the first case, a short distance to the
left of Wx; and in the second case, a short dis-

aberrant Ac conditions were present, as follows:
(1) Ac was no longer present in chromosome 9,
but was carried instead by another chromosome.
(2) Two Ac factors were present, one at the given
location in chromosome 9, and one carried by
another chromosome of the complement. (3) The
Ac factor was unchanged in its location but showed
an altered action; in a single dose it could be
equivalent to the double dose before the altera-
tion occurred, or could show an increased but
not doubled dosage action, or a decreased dosage
action.

The behavior of Ac was also studied in plants
in which the Ac factor was present at an allelic
position in each chromosome 9. Again, it could
be determined that in a few gametes produced by
such plants the above-described aberrant con-
ditions with respect to Ac position and action
were present. In addition, a few gametes were
formed that had no Ac factor at all.

From what has been said about both Ds and
Ac, it is apparent that with respect to inheritance
26 BARBARA McCLINTOCK

behavior they are much alike. The same questions
concerning mode of appearance in new locations
in the chromosome complement apply to Ac as to
Ds. With this relationship in mind, we may now
return to further considerations of Ds. It should
be emphasized again, however, that the described
events occurring at Ds, wherever it may be located
in the chromosome complement, depend upon the
presence of an Ac factor in the nucleus, regard-
less of where this latter factor is located in the
chromosome complement. New positions of Ds
activity arise only when 4c is also present in the
nucleus, and, again, regardless of where Ae may
be located. In addition, any one altered state of
Ac—for example, an altered dosage action—affects
Ds wherever it may be located, and in exactly
the same manner for every Ds, regardless of its
state. In other words, it is the state and the
dose of Ac that control just when and where Ds
events will occur, and it is the state of a partic-
ular Ds that controls the relative frequency of
any one type of event that occurs at Ds.

Tue ORIGIN AND BEHAVIOR OF cm-l

In the discussion of the appearance of Ds at
new locations, the question was raised whether
or not this involves the transposition of a material
substance from one location to another. The

question applies equally to Ac. If no material —

substance is transposed, a serious problem is
presented regarding the basic action of any known
genetic unit or factor that has been assigned to
a particular locus in a particular chromosome.
Ac clearly produces an obvious, measurable,
phenotypic response, wherever it may be located.
It shows dosage action, mendelian inheritance,
and linkage behavior of the expected type, in any
location—with the exception, already mentioned,
of a few transpositions, changes in state, and
losses of Ac. It might be considered that Ds and
Ac represent forms of altered chromosome organ-
ization producing somewhat similar effects in
each case, much like the Minutes in Drosophila.
The evidence now to be presented, however,
makes that assumption unlikely. This evidence
considers the origin and the behavior of many
different mutable loci. To begin this part of the
discussion, we may consider the origin of mutable
cm-1, the first-detected mutable c locus that arose
in a chromosome carrying a normal-behaving C
locus (C, aleurone color; c, recessive allele,
colorless aleurone).

The presence of an alteration at the known
locus of C, which produced e™-1, was detected
probably within a few nuclear generations after

it occurred. It was present only in one kernel
among approximately 4,000 examined that had
come from a cross of a single plant, used as a
male, to 12 genetically similar female plants.
All the other kernels on these ears gave the ex-
pected types of phenotypic expression. The
male parent carried in both chromosomes 9 the
genetic markers Yg, C, Sh, wx, and Ds (standard
location). It also had one Ac factor, not linked
to these markers. The female parents carried
the stable recessive yg, c, sh, and wx. No Ds
was present in their chromosomes 9, and also
no Ac factor was present in these plants. The
types of kernels to be expected from such a cross,
and their relative frequencies, are the same as
those shown in Figure 16. Approximately half
the kernels should show the C Sh wx phenotype,
with no variegation for the recessive characters
since no Ac factor is present. The other half
should carry Ac and thus be variegated. Sectors
showing the c sh phenotype should be present.
In these crosses, the expected classes of ker-
nels appeared with the exception of one kernel.
Instead of showing colorless areas in a colored
background (resulting from losses of C following
breakage events at Ds), this exceptional kernel
showed a colorless background in which colored
areas were present. The plant derived from this
kernel was tested in various ways in order to
determine the reason for this unexpected type
of variegation. The early tests indicated, and
subsequent tests proved, that mutations were
occurring from the recessive c, to the dominant
C, and that the mutable condition had arisen in
one of the chromosomes 9 contributed by the Ac-
carrying male parent plant.

Ds-type activity was also present in the chro-
mosome carrying the new mutable c locus. The
location of this activity was no longer to the
right of wx, as would be expected since this was
its location in both chromosomes 9 of the male
parent plant. The-new location was inseparable
from that of the mutable c. All the recognizable
breakage events associated with Ds-type activity
now happened at this new location. In addition,
mutations to C occurred. It was soon discovered
that the mutations to C would appear only if Ac
were also present in the nucleus, and that the
time of occurrence of thesé mutations was con-
trolled by the state and dose of Ac in precisely
the same way that the state and dose of Ac con-
trols Ds breakage events wherever Ds may be
located. If Ac were absent, neither mutations to
C nor Ds-type breakage events would occur.
Thus when Ac is absent, the behavior of em—*
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 27

is equivalent to the previously known recessive
c. If, however, Ac is again introduced into the
nucleus by appropriate crosses, the potential
mutability of this recessive is realized, for then
mutations to C occur. The. previously known re-
cessive c, used for many years in genetic studies,
is unaffected by the presence of Ac; and it re-
mains stable, nonmutable, when present in nuclei
in which ¢™-! is also present and undergoing
mutations.

In considering the mode of origin of ¢™-! and
its behavior, the following points may be reviewed:
(1) the appearance of a new recessive that is
mutable; (2) its derivation from a normal dominant
C, which is nonmutable; (3) its appearance in a
single gamete of a plant carrying Ds and Ac;
(4) Ac control of the new mutable condition in
exactly the same manner as Ds is Ac-controlled;
(5) Ds-type chromosome breakage events also
occurring at this mutable locus; and (6) disap-
pearance of Ds from its former location in the
same chromosome that carries the new mutable
locus. This series of coincidences is striking
enough in itself to command consideration of the
possibility that this mutable recessive originated
by transposition of Ds to the locus of the normal
C factor. It is immediately apparent that, if this
is true, the transposition of Ds from its former
location to the new location created a condition
that affects the formation of pigment in the aleu-
rone layer; for no pigment is formed until some
event occurs at this locus, and only when Ac is
present. Previous tests have shown that the
same c phenotypic expression can also arise if
the tissues of the endosperm are homozygous
deficient for the segment of chromatin carrying
C (McClintock, unpublished). This might suggest
that the presence of Ds has inhibited the normal
action of the chromatin materials at the C locus.
A final and most significant argument for the
origin of cm-1 by a transposition of Ds to the C
locus is derived from the fact that a mutation to
C is associated with the loss of any further
recognizable Ds events at the immediate loca-
tion of C, It is apparent, therefore, that the
mutation-producing event is associated with one
involving Ds at this locus. All the evidence is
consistent with the assumption that c™-! arose
by transposition of Ds to the locus of C, thereby
inhibiting its action, and that removal of Ds is
associated with removal of this inhibitory effect.
The restored activity at the C locus is permanent,
and its subsequent behavior resembles that pres-
ent before Ds activity appeared at the locus.
Ac no longer has any effect on its action and

behavior, just as it had no effect before Ds ap-
peared at the C locus to give rise to cm-l,

ADDITIONAL Ds-INITIATE D MUTABLE Loc!

Simple coincidence rather than a relationship
with Ds might still be claimed for the origin and
behavior of cm-1 if it were the only case of such
origin and behavior. Two other cases, similar
to cm-1 and involving another marked locus, have
appeared independently in the Ds- and Ac-carrying
plants. Both involve the locus in chromosome
9 associated with the bronze phenotype (see
Fig. 2). These two independent cases have been
designated bzm-1! and bzm-2; |The description
of the types of events occurring at cm-1 may be
applied also to bzm-1 and bzm-2, Ds-type break-
age events occur at the mutable locus, as well
as mutations from bz to an apparently full Bz
expression. Both the Ds-type breakage events
and the mutations to Bz are Ac-controlled; for if
Ac is absent, neither will occur. In these cases
also, the time when mutations to Bz or chromo-
some breaks occur is under the control of the
state and the dose of Ac. Again, as with C,
previous investigations had shown that a homozy-
gous deficiency of the segment of chromosome in-
cluding the Bz locus will reproduce the known re-
cessive phenotype, bz (McClintock, unpublished).

That the presence of Ds close to a marked
locus in a chromosome may result in frequent
changes in the phenotypic expression of the
marker has been indicated. Two independent
cases of transposition of Ds from its standard
location to a position near and to the left of Sh
have been studied (see Fig. 2). In these two
similar cases, less than one-half of one per cent
crossing over occurs between Ds and Sh. In
both cases, however, many gametes are produced
that carry a ‘‘spontaneous”’ mutation from SA to
sh. These mutations occur only in those chromo-
somes carrying Ds immediately to the left of SA,
and only in plants that also have Ac. If Ac is
absent, no such mutations appear. Ac-controlled
events, therefore, in each of these two cases
where Ds is near and to the left of Sh, are re-
sponsible for this high frequency of mutation to
sh. When such a mutation occurs, Ds is not al-
ways lost to the chromatid; it is sometimes still
present between C and Bz. Some of these
‘‘mutations’’ may prove to be newly arisen muta-
ble sh loci, but the tests for this mutability have
not been concluded.

Knowledge gained from the cases reviewed
has led to the conclusion that the appearance of
Ds at or close to the focus of a known genetic
28 BARBARA McCLINTOCK

factor can give rise to frequent changes in the
action of the factor. The initial change is to an
action resembling that of the known recessive
allele. In the cases of c”-?, bzm-1, andbz™-?,
a subsequent alteration produces a return to
apparently full dominant expression of the factor.
This common type of mutational expression in
these cases is of considerable significance.
Its importance will become evident in the dis-
cussion of the behavior of other newly arisen
mutable loci.

ORIGIN AND BEHAVIOR OF ONE CLASS OF
AuTONOMOUS MUTABLE LOCI

The behavior of some other types of mutable
loci may now be considered for the additional
and important knowledge they have contributed
to an understanding, of the basic processes in-
volved.’ One of them is called mutable luteus.
The luteus character is distinguished by a yel-
lowish chlorophyll expression.’ This mutable
luteus first appeared in the progeny derived from
self-pollination of one of the original plants that
had undergone the breakage-fusion-bridge cycle
in early development. It resulted from some
alteration at a normal locus, but the position of
the locus in the chromosome complement was not
known. This mutable locus is characterized,
first, by its autonomous behavior. © It required no
recognizable, separate activator factor in order
to undergo the mutation phenomenon. The muta-
tions are registered in sectors of a plant as
changes in the amount of chlorophyll that is pro-
duced. Alleles arise from germinal mutations.
They are characterized by various quantitative
grades of chlorophyl! expression. These alleles,
in turn, need not be stable; some of them may
mutate to give higher or lower levels of chloro-
phyl] expression. Even an allele apparently

producing the full dominant expression may be |

unstable fpr it may mutate to or towards the
lowest expression, which is luteus.

In studying one aspect of the behavior of this
mutable luteus locus, a number of sister plants
in one culture were all self-pollinated. On a re-
sulting ear of one, and only one, of these plants,

the presence of a new mutable locus was revealed. |

The mutability, registered in some of the kernels
on this ear, involved a factor associated with the
formation of pigment in the aleurone layer." Color-
less kernels were present in which mutations to
color occurred. None of the ears produced by the
sister plants showed the presence of such a
mutable factor; all kernels on these ears had the
full aleurone color. Further study of the plants

derived from the variegated kernels on the aber-
rant ear showed that a mutable condition had
arisen at a previously known locus, the A, locus
in chromosome 5 (A,, aleurone and plant color;
a,, recessive allele colorless aleurone, altered
plant color). Before the appearance of the mr
table condition in this one plant, the 4, locus
had given the normal dominant expression in both
parent plants. It had shown no indication what~-
soever of any instability. One parent had con-
tributed mutable luteus to this culture. The mu-
table luteus locus, however, was not linked with
the A, locus: It should be emphasized that this
newly arisen mutable a, behaved in many respects
like mutable luteus. Jt was autonomous; and
quantitative alleles were produced, some of
which, in turn, were mutable.

Another mutable locus arose in a culture de-
tived from one having mutable luteus. It first
appeared as a single aberrant kernel on one ear.
This ear was produced from a cross in which a
plant carrying mutable luteus was used as a fe
male parent. The male parent was homozygous
for the stable recessive a, (A,, aleurone and
plant color, located in chromosome 3; 41, reces-
sive allele, colorless aleurone and altered an-
thocyanin pigments in the plant.) This single
aberrant kernel exhibited variegation. Sectors
of colored aleurone appeared in an otherwise
colorless kernel. Tests were initiated with the
plant derived from this kernel, and continued
with the subsequent progeny. From these tests
it was learned that the aberrant kernel carried a
newly arisen mutable a, locus, designated a7"),
whose general behavior resembled that shown by
mutable luteus. It was autonomous; it produced
a series of alleles showing various grades of
quantitative expression of anthocyanin, in both
the aleurone and the plant; and of these alleles,
in turn, some were unstable, mutating to give
higher or lower levels of quantitative expression
of the anthocyanin pigments in both aleurone and
plant. Other important aspects of the mutation
phenomenon at this locus will be considered
later.

In the discussion of the Ac-controlled mutable
loci, c”"7, bz™! and bz™-2 that arose in Ds-
Ac carrying plants, it was emphasized that the
types of mutational response were similar. Here
also, the mutational expressions of the two muta-
ble loci that have arisen in the mutable luteus
stocks are much alike, and they resemble that
shown by mutable luteus itself. One further ex-
ample of related mutable loci will be given. It
also shows the similarity of behavior of the newly
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 29

arisen mutable locus to the one already present
in the plant. The direction this discussion is
taking may now be apparent. It is towards the
conclusion that the type of mutation occurring
at a locus is a function of the type of chromatin
material that is present at the locus or is trans~
posed to it, and does not involve changes in the
components of the genes themselves. Rather,
it is this chromatin that functions to control how
the genic material may operate in the nuclear
system. With this in mind, a third example of
related origins and behaviors of mutable loci
may now be considered.

ORIGIN AND BEHAVIOR OF c?™*2 AND wx!
Two RELATED MUTABLE Loci

The progeny derived from self-pollination of
another one of the original plants that had under-
gone the breakage-fusion-bridge cycle in early
development was grown, and a number of these
plants again self-pollinated. On a resulting ear
of one of these plants, a-new mutable locus was
recognized. The factor involved was again asso-
ciated with the production of pigment in the aleu-
rone layer. Some of the kernels on this ear
showed colored areas in a colorless background.
Beginning with the plants derived from these
kernels, a study was made of the condition re-
sponsible for the variegation. This proved to be
due to another new mutable locus, and involved
the previously discussed C locus in chromosome
9. This locus in the parent plant and in the
sister plants of the culture, gave the normal
dominant C expression. The new mutable con-
dition was designated c™-*, because it was the
second case that appeared in this study. The
types of mutation that arise from events at c™"?
are strikingly different from those shown by
e™1, A series of alleles, as expressed by quan-
titative grades of pigment formation associated
with the production of at least two different
precursor-type diffusible substances, is produced
by mutations at c”-?, The intermediate alleles
are not always stable, for some of them, in turn,
can mutate to alleles showing higher or lower
grades of color expression.

In the course of the study of c™?) a number
of crosses were made, using pollen of plants
homozygous for c™-? and carrying a normal
dominant Wx factor in each chromosome 9, on
silks of plants carrying a stable recessive c and
a stable recessive wx in both chromosomes 9
(see Fig. 2). A single aberrant kernel appeared
on one of the ears resulting from this type of
cross. It showed mutations to C and of the c™°?

type, and, in addition, mutations from the wx to
and towards the Wx phenotype. A plant was ob-
tained from the kernel and a study commenced to
determine the nature of this instability expression.
It proved to be a new mutable wx, and was desig-
nated wx™-1, The tests showed that it had arisen
in the male parent plant, which carried c”-? and
normal dominant Wx in each chromosome 9. It
was present, however, in only one of the many
tested male gametes of this plant. The pattern
of mutational behavior of wx™-! strikingly re-
sembled that shown by c™-2. A series of quan-
titative alleles was produced by mutations of
wx™-l as registered by the amount of amylose
starch produced. These alleles, in turn, could
mutate to give greater or lesser amounts of amy-
lose starch. Another endosperm character also
was affected by some of the mutations of wx™-!,
This was expressed by an altered growth of the
endosperm tissue, and accompanied some but not
all of the mutations to the intermediate alleles,
appearing particularly often in association with
a mutation to one of the lower alleles. This
accompanying mutation behaved as a dominant
or a semidominant.

It is of particular significance, in comparison
of the behavior of the series of mutable loci
c™1, bz™-1, and bz™-? with that of the series
c™ 2 and wx™-!, that the members of both series
are controlled by the very same Ac factor—wher-
ever it may be located—and in precisely the
same manner, with respect to time and place of
occurrence of mutations. When Ac is absent,
no mutations occur in either series. In the latter
series, mutations of the intermediate alleles
also occur, but only when Ac is present. ‘Because
of this, it has been possible to isolate a series
of quantitative alleles of C, and also a series
of quantitative alleles of Wx, that are stable.
Stability is maintained when Ac is removed. The
percentages of amylose starch in the endosperm,
produced by a number of the alleles arising from
mutations at wx! and freed of Ac, have been
determined in terms of single, double, and triple
doses of the particular allele. (Note: The writer
is grateful to Dr. G. F. Sprague and to Dr. B.
Brimhall, of Iowa State College, and also to Dr.
C. O. Beckmann, of Columbia University, and
Dr. R. Sager, of the Rockefeller Institute for
Medical Research, for their chemical analyses
of the amylose content produced by some of
these mutants.) The preliminary tests suggest
that alleles, falling into an almost continuous
series with respect te the quantity of amylose
starch they produce, may be obtained. It should
30 BARBARA McCLINTOCK

be mentioned here that, as with C and Bz, pre-
vious investigation had shown that a tissue
homozygous-deficient for the Wx locus will give
the known recessive wx expression (McClintock,
unpublished).

In order to compare the action of Ac on the
members of these two series of Ac controlled
mutable loci, crosses were made combining sev-
eral of them in a single plant so that they might
be present together in the nuclei of a tissue.
By this means, it was possible to determine that
the mutations at these various mutable loci arise
as a function of the state and dose of Ac, irre-
spective of which mutable locus is involved or
how many such loci from the same series or from
the two different series are present in the nuclei
of an individual plant.

In further comparison of the behavior of the
different Ac-controlled mutable loci, one very
significant correlation may now be given. It is
known that all these loci show one other common
characteristic. At all of them, some chromosome-
break-inducing events occur, but only when Ac
is also present in the nucleus and only at those
times in the development of a tissue where mu-
tations leading to changes in expression of the
respective phenotypic character are also occurring.
Again, it has been determined that such breaks
may occur at the locus of Ac itself. The con-
clusion that the mutation-producing events in
these two series of related mutable loci, and
also at Ac itself, are associated with such a
chromosome-break-inducing mechanism is dif-
ficult to avoid. This relationship will be ex-
plored after consideration has been given to a
comparison of the types of mutability that may
arise at any one known locus in a chromosome.

The descriptions of mutable loci given so far
in this discussion have shown that the type of
mutability and the mode of its control are not
alike for all. Nevertheless, there appear to be
classes of mutable loci, the members of which
show similar types of changes in phenotypic
expression—that is, of mutations—and similar
types of control of these mutations.
necessary to indicate the extent to which various
types of mutability expression may arise at any
one particular locus. For this purpose the Wx,
the C, and the A, loci will be chosen as examples.

CoMPARISONS OF TYPES OF MUTABILITY ~
ARISING AT ANY ONE Locus

a. The Wx locus

Six independent mutable conditions are known
for the Wx locus. Five of them have arisen during

It is now

the present study. One, wa) has been con-
sidered above. It is Ac-controlled, and produces
quantitative alleles that may be unstable when
Ac is present but are stable when Ac is absent
from the nucleus. Mutations at this locus also
give rise to an endosperm-growth-altering factor
that is dominant in expression. The second
mutable wx, wx™-?, arose from a previously stable
recessive wx carried in genetic stocks for many
years. This mutable condition first appeared in a
chromosome 9 in which a complex chromosomal
rearrangement was present. Its mutations are
expressed by different quantitative grades of the
Wx phenotype. In the endosperm tissues, the
sector produced by a cell in which a mutation has
occurred is always markedly distorted in growth-
type. The third mutable condition at this locus,
wxm-3, originated in a plant carrying a normal
dominant Wx and also several mutable loci. It
is autonomous, in that no separate activator
factor is required for mutations to be expressed.
It almost always mutates to give the full Wx
phenotypic expression. The derived mutant giving
the dominant expression is also mutable, for it
produces mutants giving the full recessive ex-
pression, that is, wx. This recessive, in turn,
may mutate again to give the full dominant ex-
pression. No altered growth conditions in the
endosperm tissue accompany any of these muta-
tions. The fourth case was recognized by a
sudden change in the behavior of a previously
normal Wx locus. It shows mutations producing
various grades of quantitative expression between
the full dominant and the full recessive. It is
autonomous, and no alterations in growth con-
ditions accompany the mutations. Another case,
somewhat similar to the last, has recently arisen.
It produces alleles giving various lowered ex-
pressions of the Wx phenotype, and-appears to be
autonomous although the information on its be-
havior is too incomplete to allow a full descrip-
tion. The sixth case is one that has been in-
vestigated by Sager (1951). It is autonomous,
and gives quantitative grades of expression in
the endosperm; but the germinal mutations that
have been studied all give full or nearly full Wx
expression, and the mutants are stable. No al-
tered growth conditions appear to accompany
these mutations. .

Genetic analyses have indicated that all these
various mutational changes occur at this one
locus in chromosome 9, and yet all show a dif-
ferent kind of mutational behavior. It is evident
that each arose in association with a particular
type of alteration at the locus, and that different
mutation-controlling mechanisms can be involved.
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 31

This is especially well illustrated by a comparison
of the types of mutations produced by wx™*! and
wxm™-3 and of their controlling mechanisms.

b. The C locus

The contrasts in the kinds of phenotypic ex-
pression produced by mutations at c”-! and em-2
have been discussed above. Although several
other independent expressions of mutability at
this C locus have also arisen, the study of them
is too incomplete to allow detailed comparisons
to be made. With respect to this locus in chro-
mosome 9 it is necessary to mention, however,
that in the cultures having mutable loci a herit-
able factor carried by a chromosome other than
9 has appeared on several occasions. The pres-
ence of such a factor results in the production
of pigment in the aleurone when the endosperm
is homozygous for the well-investigated stable
recessive c, used for many years in genetic in-
vestigations. The pattern of pigment formation
differs markedly from that produced by mutations
at c™-1 or c”"*, It resembles that associated
with the factor Bh (Blotch), previously studied
by R. A. Emerson (1921) and Rhoades (1945b).
To complete the discussion of the series at this
locus, it may be mentioned that a mutable con-
dition has also arisen involving the expression
of /, an allele of C. Changes in the degree of
inhibitory action of 7 occur as a consequence of
such mutations.

c. The A, locus

A study of changes at the locus of A, has con-
tributed some very important information regarding
the origin and behavior of mutable loci. For a
number of years, a type of control of mutability
of the recessive, a,, has been known. A dominant
factor, called Dotted (Dt), provokes mutability
at the a, locus (Rhoades, 1936, 1938, 1941,
19452). In many respects, Dt is comparable to
Ac. It is an activator, for it produces mutations
at a,, just as Ac produces mutations at em-1,
cm-2) bzmol bzm-2) and wx™-1, Moreover, when
Dt is absent no mutations occur at a; the a
locus then gives a stable recessive phenotype.
In the presence of Dt, mutations occur at a; to
give mainly the higher alleles of the A, phenotypic
expression. The time of occurrence of visible
mutations at a, is usually late in the develop-
ment of the plant or the endosperm tissues; and
they occur in only some of the cells, This re-
sults in the presence of dots of the A, phenotype.
The De factor has been located by Rhoades in
the knob region terminating the short arm of chro-
mosome 9, Det and Ac appear not to be the same

activator, as plants and endosperm tissues that
are homozygous for the recessive a, have not
shown the dotted-type mutations to A, in the
presence of Ac.

In the early period of this investigation of
newly arisen mutable loci, the unexpected ap-
pearance of modifications in the knobs, and in
the other chromosome elements previously men-
tioned, of plants that had undergone the breakage-
fusion-bridge cycle in their early development,
suggested that disturbances in these elements
might have been responsible for the initial burst
of mutable loci, including the origin of Ac itself.
It was suspected, therefore, that this cycle might
induce alterations in the heterochromatic ele-
ments that could initiate a De factor as they may
have originated the Ac factor. Once initiated,
this factor would activate a, to undergo altera-
tions in somatic cells leading to A-type ex-
pression. The most direct way to induce changes
in the heterochromatic elements was considered
to be the breakage-fusion-bridge cycle itself.
By subjecting tissues to this cycle during their
developmental periods, and then examining the
matured tissue, this hypothesis could be tested.
A preliminary experiment, designed to test for
production of mutations of a, to A, by the breakage-
fusion-bridge cycle as an inductor, was performed
in 1946. The experiment was repeated in 1950
on a much larger scale. Because this experi-
ment was of particular significance in revealing
the mode of origin of mutable conditions, and
because it provided evidence about the relation
of chromosome organization to genic expression
and its control, the details will be given.

The silks of plants homozygous for a, and
carrying no Dt factor received pollen from plants
of similar constitution with respect to a, and Dt.
The pollen parents carried one chromosome 9
with a duplicated segment of the short arm. The
homologous chromosome 9 was deficient for a
terminal segment of the short arm. Newly broken
ends of chromosome 9 were produced in some
meiotic cells, as diagrammed in Figure 1. Pollen
grains of these plants carried either: (1) a de-
ficient chromosome 9, which did not function in
pollen-tube growth, (2) a chromosome 9 with a
full duplication of the short arm—that is, the
homologous chromosome 9, or (3) a chromosome
9 with a newly broken end. Among this last type,
duplications or deficiencies of the short arm
were present. Those carrying an extensive de-
ficiency were nonfunctional but those carrying
a relatively short duplication were better able
to compete in functioning than those carrying the
full duplication of the short arm. Thus the
32 BARBARA McCLINTOCK

majority of the functioning pollen grains of these
plants carried a newly broken end of the short
arm of chromosome 9 in their nuclei. These chro-
mosomes had undergone the breakage-fusiom
bridge cycle since the meiotic anaphase and con~
tinued to do so after being incorporated into the
primary endosperm nucleus. Either before fer
tilization or during the development of the kernel,
the breakage-fusion-bridge cycle might produce
alterations in heterochromatic elements and some
of them might include an alteration that would
recreate the condition associated with De action.
Mutations at the a, locus to give the A, phenotype
could subsequently appear in the descendant
cells. If this should occur early in development
of the endosperm, a sector with dots of the A,
phenotype would be produced. Examination of
95 ears resulting from the preliminary test con-
ducted in 1946 revealed A, dots on 15 kernels
from 14 different ears. In five of these kernels,
more than one A, dot was present, and in one
restricted region of the kernel in each case.
The number of kernels with mutations to A, in
this trial experiment was lower than anticipated,
and the experiment was not expanded the follow-
ing year. Later,as the probable relation between
the origins of mutability and the alterations in-
duced inthe knobs or other chromosome elements
by the breakage-fusion-bridge cycle became more
clearly apparent, the same experiment was con-
ducted in 1950 on a much larger scale. The re-
sults were rewarding, for now many kernels were
obtained that had one or more A, dots. One hun-
dred and twenty such kernels appeared in this
second trial, and 24 of them had more than one
A, spot. One of these kernels had 84 A, spots,
distributed rather evenly over the kernel. In the
other 23, the spots were not distributed at ran-
dom over the aleurone layer but were restricted
to well-defined sectors. In none of these kernels
did any large areas of the A, phenotype appear.
In all cases, the time of mutation, the pattern
of mutation, and the type of mutation were much
like those produced when the known Dt factor is
present in endosperms homozygous for a. It was
obvious that in each case the initia] alteration
had occurred in the ancestor cell that produced
the dotted sector. This initial event was re-
sponsible for mutations that occurred at a, in
some cells during. the subsequent development
of the endosperm. The observed mutations at
a, therefore, were not produced directly by the
breakage-fusion-bridge cycle but arose second-
arily, as a consequence of an event that altered

some patticular component in the nucleus. It

was the alteration of this component that was
responsible for the subsequent mutations at the
a, locus. And this initial alteration was one that
smitated the effect produced when the known Dt
factor is present. It is difficult to avoid the con-
clusion that a new De-like factor has been pro-
duced in each such case, and that it was created
by some event associated with the breakage~
fusion-bridge cycle. Unfortunately, the plant
grown from the one kernel having 84 dots dis-
tributed over the whole aleurone layer did not
show any mutations to A,, nor did mutations
appear in the kernels when this plant was crossed
to plants homozygous for a1 The effective al-
teration probably was present in only one of the
two sperms carried in the pollen grain. Because
the break in chromosome 9 that initiated the
breakage-fusion-bridge cycle was produced at
the meiotic anaphase, the event giving tise to
the Dt-like factor would have had to occur in the
subsequent microspore division in order to be
incorporated in the two sperm nuclei. An even
larger experiment of this same type must be con-
ducted if such a case is to be obtained. It
should be mentioned in this connection that the
size of the sectors within which A, spots ap-
peared graded from large to small, the smaller
sectors being most frequent. Also, about three-
fifths of the kernels showing mutations to A,
had only one A, spot. These frequencies are to
be expected if the creation of a Di-like factor
is a consequence of an event, associated with
the breakage-fusion-bridge cycle, that has a
probability of occurrence in a limited number of
mitotic cycles. In order to indicate why the dotted
pattern of mutations to A, is to be anticipated,
rather than any other, it is necessary to review
the origin of the previously discovered Dt — a1
mutable condition.

The Dt — a, mutable condition first appeared
on one ear after self-pollination of a plant be-
longing to the commercial yariety known as
Black Mexican Sweet Corn. This variety is
homozygous for A, The recessive a; in this
case represented a new mutation from A,, and
was associated with the appearance of Dz. The
original a, mutant, known for many years and
used in genetic studies, had originally been
found to be present in a commercial variety of
maize. Both’ a, mutants responded in much the
same manner when Dt was present in the nucleus.
In both cases, the dotted mutation pattem was
produced in the presence of Dt. The states of
the two a, mutants thus appeared to be alike.
This suggests that the older a, mutant may
CHROMOSOME ORGANIZA TION AND GENIC EXPRESSION 33

have been produced by a mechanism similar to
that responsible for the origin of the newer a,
mutant. A Dt factor may have arisen at the same
time but subsequently been lost from the com-
mercial variety during its propagation, leaving
an apparently stable a, mutant. The change at
C that produced c™-! would have behaved quite
comparably had Ac been absent from the nuclei
in the initial gameté carrying c™-1, or had it
been removed by crossing before the change
at this locus had been detected. If the mutation
had been discovered several generations after
its origin, and if Ac had been removed by a
previous cross, it would have appeared to be a
newly arisen, stable, recessive c. Only after an
incidental cross to a plant carrying Ac would

its potential mutability have been revealed. It

is possible, therefore, that many known reces-
sives may prove to be potentially mutable.

The essential similarity of the Dt — a, system
to the Ac — cm-1, etc. system is also expressed
in the changes in state of a, that may occur in
the presence of Dt, Such changes in state of a,
have recently been described by Nuffer (1951).
They are recognized individually by marked de-
partures in frequency of visible mutations, in
types of mutation and in time of occurrence of
these mutations. The types of different pheno-
typic expression produced by mutations at al-
tered states of a, are much the same as those
produced by a mutable a; locus that has appeared
in the Cold Spring Harbor cultures. This new
mutable a, locus, called a,™-/, differs from the
mutable a, studied by Nuffer in that it is auton-
omous and does not require Det for mutability to
be expressed. In this respect, mutability at the
A, locus behaves like that at the C and the Wx
loci, for both autonomous and activator-controlled
mutable conditions may arise.

The origin of a,™-1, in a culture carrying mu-
table luteus, was described previously. It is
autonomous, and produces a series of quantita-
tive alleles, many of which are unstable in that
they may mutate to give higher or lower levels of
quantitative expression. Difference in degree of
quantitative expression is only one of the con-
sequences of mutations occurring at a,7-1, how-
ever. The diversity of phenotypic changes aris-
ing from these mutations is so great that an ade-
quate analysis of all the observed types is a
large task. They are distinguished not only by
quantitative but also by qualitative differences
in the anthocyanin pigments formed. Diversity
is shown in other respects. For example, some
of the mutations giving pale aleurone color are

related to changes involving the rate of a partic-
ular reaction responsible for pigment formation.
Others appear to be related to the absolute amount
of pigment that may be produced, regardless of
a time factor. This becomes evident when com-
parisons are made of pigment-forming capacities
in plants arising from kernels carrying different
mutants of a,™-1, each producing a pale color
in the aleurone of the kernel. In some cases,
such plants are palein their expression of antho-
cyanin color throughout their lives. Others are
pale in anthocyanin color up to the time of an-
thesis, when growth of the plant terminates; but
in the six or seven following weeks, as the plants
mature their ears, the anthocyanin color gradually
deepens, becoming intensely dark by the time
the ears are mature. The kemels derived from
both these two types of plants, however, may be
equally pale. The fact that pigment forms late
in the development of the kemel, and dehydration
of the tissues occurs shortly thereafter, may ex-
plain this similarity in color of the kernels.

Other types of mutation occur at a,7-!. Some
produce sectors of deep color that are rimmed by
areas in which the color gradually fades off to
colorless, as if a diffusible substance associated
with pigment formation had been produced in
excess in the mutant sector. The area of dif-
fusion may be very extensive for some mutations,
and only slight for others, whereas still other
mutations give rise to no such diffusion areas
at all. Some of the mutations that result in
strong A, pigmentation are associated with
failure of development or degeneration of some
of the aleurone cells within the mutated sector.

Besides the mutational changes at a,7-! that
affect the type and amount of pigment formation,
a number of other changes occur which affect
the subsequent behavior of a,7-!. These altera-
tions are termed changes in state, since they
affect not only the times when pigment-forming
mutations will occur at the locus in future plant
and endosperm cells but also the kinds and fre-
quencies of such mutations, and their distribu-
tions and their sequences in the development of
the tissues,

Any interpretations that attempt to explain the
primary action of a specific locus in a chromo-
some, and how this action may be changed, must
take into consideration the facts just enumerated
concerning the behavior of this a,™-/ locus. It
is not reasonable to regard such changes in ex-

_ pression and action as being produced by changes

in a single gene—that is, according to the usu-
ally accepted concepts of the gene that have
34 BARBARA McCLINTOCK

been developed. The evidence suggests, rather,
that the observed changes result either from
alterations at a locus that has many individual
components or from alterations at the locus af-
fecting its relationship to other loci in the chro-
mosome complement. If the latter is true, a
combination of loci functions as an organized
unit in the production of pigmentation. If such
functional organizations exist within the nucleus
—and it is reasonable to assume they do--then
the large numbers of alleles known to arise at
certain loci need not express altered genic action
at the identified locus. Rather, any one altera-
tion may affect the action of the organized nu-
clear unit as a whole. The mode of functioning
of various other loci concerned may thereby be
modified. In other words, the numerous different
phenotypic expressions attributable to changes
at one locus need not be related, in each case,
to changes in the genic components at the locus,
but rather to changes in the mechanism of asso~
ciation and interaction of a number of individual
chromosome components with which the factor
or factors at the locus are associated. Accord-
ing to this view, it is organized nuclear systems
that function as units at any one time in develop-
ment. In this connection it may be repeated that
at a,™-!, and also at other mutable loci, many
of the alterations observed represent changes in
the potential for patterns of genic action during
development (changes in state). Thus a pattern-
controlling mechanism is being altered. If partic-
ular nuclear components are formed into organized
functional nuclear units, the evidence would sug-
gest that this may happen only at prescribed times
in the development of an organism. In this event
it may readily be seen that changes in pattern-
controlling mechanisms would serve as a primary
source of potential variability of genic expres-
sion without requiring any changes in the genes
themselves.

A few morg pertinent facts about the A, locus
may now be mentioned. Mutability has arisen
at this locus independently on a number of dif-
ferent occasions. Several cases have recently
appeared in the stocks having Ds and Ac. Anal-
ysis of these cases has not proceeded to a stage
where a complete description of their behavior
may be given. Both Laughnan (1950) and Rhoades
(1950) have found new cases of instability at
this locus. Several have appeared in plants de-
rived from kernels that had been aged for some
time (Rhoades, 1950), Such aging is known to
give rise to chromosomal aberrations as well as
mutations; the observed instability may be an

expression of one such structural change.
Laughnan (1949, 1951) has shown that mutations
at the A, locus may be associated with the
mechanism of crossing over, suggesting again
that mutations arise from structural changes at
the locus. The crossover studies may elucidate —
the nature of some of these changes. Not only
the cases described in this report but also others
have produced evidence converging in support of
a hypothesis that mutations originate from struc~
tural alterations in chromosome elements. The
evidence derived from a study of progressive
changes in state of c™-! has shown the close
relation between a structural change at a locus
and one that so often has been called a “gene
mutation.” This study will now be described.

SIGNIFICANCE OF CHANGES IN STATE OF A
MUTABLE Locus: SELECTED EXAMPLES

The foregoing review of the very different types
of phenotypic expression that may be produced
as a consequence of mutations arising at any one
locus, and of the relation between the origin of
a mutable condition and the type of mutations
expressed, clearly indicates the necessity for
caution in attempting to interpret the mutation
process as one associated with a ‘change in a
gene.” With respect to events occurring at Ac,
em-1, em-2, bzm-1, bzm-2, and wxm-l, it has
been established that a mechanism capable of
producing chromosomal breaks at the locus is
associated with the mutation-producing process.
It could be argued from the evidence so far pre~
sented that these cases fall into a special! cate-
gory, and that what they may indicate regarding
the mechanism of mutation at these mutable loci
may not be used to interpret the mutation process
in general. Knowledge gained from a study of
the changes in state of c™-! has shown, however,
that no line may be drawn between those events
at a locus that produce detectable chromosomal
alterations and those that give rise to mutations
but produce no readily detectable chromosomal
alterations.

The origin of c”-! by a transposition of Ds to
the normal C locus has been discussed above.
The state of Ds, when first transposed to the Cc
locus, was one that produced many detectable
chromosome breaks at this locus and few mutations
to C. When plants having this state of c7-1 were
crossed to plants that were homozygous for the
stable recessive c, the majority of the resulting
kernels showed this relationship. Some of these
kernels, however, had sectors with higher rates
of mutation to C and lower rates of detectable
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 35

chromosome - breaks. In some sectors, no chromo-
some breaks were evident; but often a high fre-
quency of mutation to C had occurred. There were
also a few kernels on these ears that had this
pattern throughout the kernel. When found, such
kernels were selected from the ears (as well as
others that showed changed mutational patterns).
The plants grown from them were again crossed
to plants homozygous for the stable recessive c.
The kemels on the resulting ears now showed
the types and frequencies of the different de-
tectable events at c”-! that had been observed
in the kernel from which the plant had arisen.
It was possible to determine in this manner that
a heritable change had occurred at c™-1,. It was
this change which was responsible for the altered
frequencies of expression of the detectable events
that subsequently occured at this locus. That
the altered response, in each isolated case,
arose from a change at c™-! and was not pro-
duced by a change at Ac, was determined by
testing the responses of these c”-! isolates with
different isolates of Ac, each having a known
type of action. Ac controlled, in each test, the
time and place of the event occurring at c™-1,
but not its type. Because, in each case, the
change in behavior of c™-! was heritable, it must
have arisen by an event that produced an altera-
tion at this locus. It is this heritable altered
condition that has been termed an altered state
of the locus,

The various altered states of c™-1, as pre-
viously mentioned, arise only when Ac is also
present in the nuclei, and only at the times in
development when mutations or chromosome
breaks may occur at c”"?, For our purposes,
the most instructive of the changed states are
those giving reciprocal frequencies of chromo-
some breaks and mutations to C. A series of
isolates, each showing a particular relation be-
tween these two events at c”"!, has been studied.
The isolates ranged from those showing no mu-
tations to C or only a very occasional one, but
having a very high frequency of detectable chro-
mosome breaks, to those showing a high fre-
quency of mutations to C and no detectable break-
age events or only an occasional one. The states
of c™-! that give high frequencies of chromosome
breaks are unstable, for other altered states may
be produced as a consequence of events at the
locus, but only, as emphasized above, when Ac
is also present in the nucleus. A particular state
of c™-! remains constant if maintained in plants
having no Ac. The state of c™-1 giving no de-
tectable chromosome breaks, and a correspond-

ingly high frequency of mutations to C, is very
stable with respect to the absence of breakage
events. Had the state of c™-! been of this type
when it first arose, there would have been no
opportunity to discover that the chromosome-
break-producing mechanism and the mutation-
producing mechanism were related. If chromosome
breaks are not exhibited by a mutable locus,
therefore, it cannot be argued that because of
this the basic mechanism producing the mutations
must be different from that known to operate at
e™1 and at the other Ac-controlled mutable loci.
The evidence obtained from this study of the
origin and subsequent behavior of-altered states
of c™-1 argues, rather, for similarities if not
identities in the basic mechanism.

Another type of change in state of c”*! should
be mentioned, although it occurs infrequently.
It is detected by a much altered expression of
the mutation at this locus that affects aleurone
color. The frequency of origin of such states
of cm-1 is so very low as to suggest that they
represent entirely new modifications at this
locus, comparable to the original inception of a
mutable condition at any locus. They may well
represent just such new inceptions, for these are
to be expected in view of the fact that very dif-
ferent types of mutational behavior are exhibited
at this same locus by c””” and c™"*, both of
which are Ac-controlled.

All the Ac-controlled mutable loci exhibit
changes in state. These are characterized by
changed relative frequencies of the different
recognizable types of mutations that occur. In
other words, as we have already seen, the types
of mutation produced in Ac-controlled mutable
loci are related to the state of the mutable locus
itself. The time and place during development
of occurrence of mutations, on the other hand,
are controlled by the state and dose of Ac; and
therefore alterations in them are related to changes
in state of Ac rather than to changes in state of
the mutable locus. The changes in state of Ac
have been described. The autonomous mutable
loci also undergo changes in state, as described
previously for the a,”"! locus. In this group,
however, the controller of the time and place of
occurrence of mutations is a component of the
locus itself. Consequently, the changes of state
that arise at these loci are reflected in changes
in the control of time and place of mutations as
well as in the type or types of mutation that may
occur, and also in the time and place of occur-
rence of each such mutation if several types
are produced. Thus there is a much more diverse
36 BARBARA McCLINTOCK

group of altered states associated with changes
at any particular locus in the autonomous group.
The general similarities between the autonomous
and the Ac-controlled mutable loci are neverthe-
less striking.

EXTENT OF INSTABILITY OF GENIC
EXPRESSION IN MAIZE

The discussion so far has mentioned a number
of different mutable conditions at known loci in
maize. In order to show that the phenomenon is
much more prevalent than the particular cases
described would indicate, some additional obser-
vations of changes in genic expression may be
discussed briefly. Not only have C, 1, Bz, Wx,
A,, and A, become mutable during the course of
these studies, but instability of other known
dominants has been noted. These are R, Pr, Yg,
Pyd, Y, and possibly B and Pl, although the evi-
dence for the last two is observational and not
genetic. Some previously unknown dominant loci
have also become unstable. These are associated
with various chlorophyll-determining factors,
endosperm-starch-controlling factors, aleurone
color factors, growth-controlling factors, etc.
Instability has arisen at the loci of some of the
known recessives such as wx, yg; and the special
case a, described previously. Instability at the
loci of the recessives y and p also appears to
have arisen on several independent occasions.
Genetic tests were not made, however, to deter-
mine the association of the instability with the
known loci. Instability arising at recessive loci
is recognized by mutations to or towards the ex-
pression of the dominant allele. It may be con-
cluded, therefore, that many -of the known reces~-
sive alleles are potentially capable of expressing
action that is characteristic of the dominant
alleles.

The expression of instability of various factors
in maize is probably far more common than has
been suspected in the past. Until recently, only
a few such cases had been reported in the litera-
ture. One of the earliest recognized was that
occurring at the p locus (pericarp and cob color,
chromosome 1), studied by R. A, Emerson and
his students (for literature citations, see Demerec,
1935) and recently being studied by Brink and
his students (1951, and personal communication)
and by Tavcar (personal communication). _ Re-
ported cases of instability at the a and the wx
loci have been mentioned previously in this dis-
cussion. In addition, Rhoades (1947) has studied
instability at the bt locus (brittle endosperm,
chromosome 5) and has reported two cases in-

volving chlorophyll characters that appeared in
the progeny of irradiated seed (Rhoades and
Dempsey, 1950). Fogel (1950) has been investi-
gating instability at the R locus (aleurone, plant,
and pericarp color, chromosome 10); and Mangels-
dorf (1948) has reported instability at the Tu locus
(Tunicate, chromosome 4). It is believed that
critical examination will uncover many such cases
in maize, and that they will involve many different
loci.

MUTABLE LOCI AND THE CONCEPT OF THE
GENE

It will be noted that use of the term gene has
been avoided in the foregoing discussion of in-
stability. This does not imply a denial of the
existence within chromosomes of units or ele-
ments having specific functions. The evidence
for such units seems clear, The gene concept
stems from studies of mutation. That heritable
changes affecting a particular reaction, or the
development of a particular character, in an organ~
ism arise repeatedly and are associated with a
change of some kind occurring at one specific
locus or within one specific region of a chromo-
some, has been established. This knowledge has
been responsible for the development of a con-
cept requiring unitary determiners. It cannot be
denied, in the face of such evidence, that certain
loci or regions in the chromosomes are associated
in some manner with certain cellular reactions or
with the. development of particular phenotypic
characters. This is not the major questionable
aspect of current gene concepts. The principal
questions relate to the mode of operation of the
components at these loci, and the nature of the
alterations that affect their constitution and their
action. Within the organized nucleus, the modes
of operation of units in the chromosomes, of what-
ever dimensions these may be, and the types of
change that may result in specific alterations in
their mode of action, are so little understood
that no truly adequate concept of the gene can
be developed until more has been discovered
about the function of the various nuclear com
ponents. The author agrees with Goldschmidt
that it is not possible to arrive at any clear under-
standing of the nature of a gene, or the nature of
a change in a gene, from mutational evidence
alone. At present, the most we know about any
“gene mutation’ is that a heritable change of
some nature has occurred ata particular locus
in a chromosome, and that any one locus is some~
how concerned with a certain chemical! reaction,
or with a certain restricted phenotypic expres-
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 37

sion, or even with the control of a complex pat-
tern of differentiation in the development of a
tissue or organ. The various types of known mu-
tation, each showing unitary inheritance, obviously
reflect various levels of control of reactions and
reaction paths. It is necessary to consider these
various widely different levels of unitary control
and how they may operate in the working nucleus,
and also to consider the nature of the changes
that can affect their operation. It is with the
nature of such heritable changes, the conditions
that induce them, and their consequences, that
this report has been concerned. Various levels
of unitary control, as witnessed by inheritance
behavior, are evident from the study. That genes
are present in the chromosomes and that they
function to produce a specific type of reactive
substance will be assumed in this discussion,
even though such a restricted assumption may
prove to be untenable. The knowledge gained
from the study of mutable loci focuses attention
on the components in the nucleus that function
to control the action of the genes in the course
of development. It is hoped that the evidence
may serve to clarify some aspects of gene action
and its control. Some of the interpretations of
the author, based on this evidence, have been
stated or implied at various points in the pre-
vious discussion, and may now be summarized.
The primary thesis states that instability
arises from alterations that do not directly alter
the genes themselves, but affect the functioning
of the genic components at or near the locus of
alteration. The particular class to which a mu-
table locus belongs is related to the particular
kind of chromatin substance that is present at or
near the genic component in the chromosome. It
is this material and the changes that occur to it
that control the types and the rates of action of
the genic components. Thus the basic mechanism
responsible for a change at a mutable locus is
considered to be one that is associated with a
structural alteration of the chromatin materials
at the locus. The mechanism that brings about
these changes is related to the mitotic cycle;
and it may involve alterations of both sister chro~
matids at the given locus. Some of these altera-
tions may immediately result in the expression of
an altered phenotype, a “‘gene mutation.”’ Others
produce modifications controlling the type of
events that will occur at the locus in future cell
and plant generations. Still others produce
changes of a more extensive type, such as dup-
lications and deficiencies of segments of chro-
matin in the vicinity of the locus. With regard

to these conclusions, the evidence presented in
the discussion of changes in state of c™-1 may

be recalled.

BEHAVIOR AND ACTION OF Ac IN CELL
AND PLANT GENERATIONS

The interpretations given above deal with the
organization and the kinds of events that occur
at genetically detectable loci in the chromosomes.
The next level of consideration deals with mode
of operation of the nucleus in controlling the
course of events during development. Do these
studies suggest a mode of operation, or at least
one component in the operative system? It is
believed that the behavior of Ac may be of im-
portance in such a consideration. With reference
to em-1, for example, it has been shown that mu-
tations to C occur at particular times and places
in development, under the control of the partic-
ular state and dose of Ac. It has also been
shown that Ac is unstable, for changes arise
affecting its dosage action, and changes also
occur in its location in the chromosome compl e~
ment. It has not been explained, however, that
the time of occurrence of these changes at dc
during development is also a function of the
particular state and dose of Ac that is present
in a tissue. With any particular Ac, and any
particular dose of this Ac, or with any combina-
tion of different isolates of Ac, the time when
these events will occur to Ac is a function of the
single or combined action of the Ac factors. An
example of this may now be given.

One plant having an Ac factor at the same
location in each homologue of a pair of chromo-
somes, and carrying /, Sh, Bz, Wx, and Ds (stand-
ard location) in both chromosomes 9, was crossed
to a number of plants homozygous for C, sf, bz,
and wx and carrying no Ds or Ac. On all the
many ears resulting from these crosses, approxi-
mately 90 per cent of the kernels were sectorial
with reference to the time of occurrence of Ds
events. Pollen from this same plant was placed
on silks of plants having other combinations of
markers in chromosome 9, and similar types of
sectorial kernels appeared. The sectoring was
produced by segregations, occurring in the earliest
nuclear divisions of the endosperm, that involved
the controller of the time of occurrence of Ds
events, that is, the Activator. The action of
Ac in these different sectors resembled that oc~
curring either (1) when no Ac is present, (2) when
a sharply decreased dose of Ac is present, or
(3) when a sharply increased dose of Ac is pres-
ent. Illustrations that will make this relation-
38 BARBARA McCLINTOCK

ship clear are shown in the photographs of Figures
17 to 22. In a definite fraction of the cases, a
chromosome break at Ds was associated with
segregations of Ac action. The mechanism re-
sponsible for this precise, somatically ocewring
segregation for Ac action was very likely the same
as that responsible for the origins of germinal
changes in action of Ac, as well as changes in
its position in the chromosome complement. Such
changes have been mentioned earlier in this re-
port. It can be seen that if this same type of
segregation occurred within some cells early
enough in the development of a plant to be in-
corporated in a microspore or megaspore nucleus,
the altered state or location of Ac, or both, would
be recoverable in the plant that subsequently re-
sulted from the functioning of the male or female
gametophyte arising from such a spore. Such an
early-timed event would allow for isolation and
subsequent study of the transpositions and
changes in state of Ac.

A study was initiated to determine the nature
of the changes that occur at Ac by an analysis
of the Ac constitutions in the gametes of plants
having an Ac factor at the same location in each
member of one pair of chromosomes, that is, in
plants homozygous for Ac in allelic positions.
In these plants the Ac factors were alike in state,
and the homozygous condition was produced by
self-pollination of plants having. one Ac factor.
Both chromosomes 9 in these plants carried the
stable factors c, sh, and Wx. No Ds was present
in these chromosomes. To test for Ac inheritance,
these plants were crossed by plants having no
Ac but carrying C, Sh, wx, and Ds in both chro-
mosomes 9 (standard locations of Ds and identical
states). If no changes had occurred to Ac, all
the kernels on the resulting ears should be varie-
gated. Similar variegation patterns, produced by
sectors showing the c, sh, and Wx phenotype,
should be present, because Ac would initiate

chromosome breaks at Ds in the C, Sh, wx, Ds-
carrying chromosome 9 contributed by the male
parent. With the exception of a few kernels, just
such conditions were realized on these ears. A
photograph of one such ear appears in Figure 23.
It will be noted that the majority of kernels on
this ear show very similar patterns of variegation.
A few kernels that differ from the majority are
completely colored, with no colorless sectors of
any size. In them, no Ds breaks at al! occurred.
A few other atypical kernels show an altered
timing of the breakage events at Ds. In them
such events occurred either much earlier or very
much later in development of the endosperm.

A study was made of the Ac constitution in
plants derived from selections of all the different
types of kernels appearing on such ears. In the
plants coming from the kernels showing altered
variegation patterns, it was necessary to deter-
mine the subsequent behavior of Ac; and in the
plants coming from the nonvariegated kernels it
was necessary to determine the presence or ab-
sence of Ac, and the presence or absence of Ds
in the C, Sh, wx-carrying chromosome. The re-
sults of this study may be summarized. In the
plants derived from the majority class of kernels,
a single Ac factor was present. Its state was
similar to that present in the parent plant (more
than 25 cases studied). The Ac constitution in
the plants derived from the nonvariegated kernels
was most instructive. In 19 plants, no Ac was
present. In 17 plants, two nonlinked Ac factors
were present. In six plants, an Ac factor in-
herited as a single unit was present, but it gave
a dose action equivalent to two doses of the Ac
factor in the parent plant. In the plants derived
from kernels that showed very late-occurring Ds
events, either two nonlinked Ac factors were
present (5 cases), or a single Ac factor was
present giving a dose action greater than that of
the Ac factor in the parent plant (3 cases).

 

FIGS. 17 to 22. Photographs of kernels illustrating the somatic segre

gations of Ae that may occur very

early in the development of a kernel. These kernels arose from the cross of plants (¢) carrying C, bz and no
Ds in each chromosome 9 and having no Ac factor, by plants carrying I, Bz, and Ds (standard location) in chro-

mosome 9 and also carrying Ac. For phenotypes exp
Figures 10 to 15.

ected from breaks at Ds, see descriptions accompanying
In Figures 17 and 18 there are 4 large sectors in each kernel: one is C bz (above in Figure

17, to right in Figure 18), one is I, non-variegated (to right in Figure 17, upper left in Figure 18), one is char-

acterized by late occurring Ds breaks, producing speckles of the C bz ge

notype (left in Figure 17, lower left

in Figure 18), and one shows that numerous Ds breaks occurred earlier in the development of the kernel (lower

segment in Figure 17, middle segment in Figure 18).

The kernel in Figure 19 has 3 sectors: one that is C bz,

one with few specks of the C bz genotype and one with many specks of the C bz genotypes. Figure 20 shows a

kernel with 3 sectors: one that is C bz, one that is wholly / Bz and one

having many specks of the C bz geno-

type. Figure 21 shows a kernel with two sectors: one with many specks of the C bz genotype and one with few

such specks. Figure 22 shows a kernel with five sectors:

a large C bz sector (lower left), a large sector having

many specks of C bz (upper), a large sector showing many larger C bz areas (upper right), a small sector with
few C bz specks (middle), and a sector of | Bz with no C bz specks (lower right),
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 39

 
40 BARBARA McCLINTOCK

From this analysis it is clear that all the aber-
rant kernels on ears of the type shown in Figure
23 were produced because of some alteration of
Ac that had occurred in cells of the parent plant.
The reason that no Ds breaks were detected in
some of these kernels is related either to the
absence of Ac in the endosperm or to the pres-
ence of a marked increase in the dose of Ac. It
will be recalled that the female parent contributed
two gametophytic nuclei to the primary endosperm
nucleus. If each nucleus carried two Ac factors,
or a single Ac factor with a double-dose action,
the endosperm would have either four Ac factors,
or two Ac factors equivalent in action to four
Ac factors. In such kernels, the high dose of Ac
so delayed the time of occurrence of Ds breaks
that none took place before the endosperm growth
had been completed.

In order to verify the analyses of Ac constitu-
tion in some of these cases, tests were continued

  
      

    
  
    
   
 

: ( ; oy

iy.

Pre

 
  
 

 
 

  
 
  
   
   
    
  

  

      
 

 

a ee nd TS
en bee
ee
+ et
Saber a tend
ePpheed Cro
ee eae
ee 4-ae
ate
"3 ue
ae
Rea Te
Lat, ange Ag

FIG. 23. Photograph of an ear derived from a plant
having two identical Ac factors located at allelic posi-
tions in an homologous pair of chromosomes, This
plant carried ¢ in each chromosome 9. The & parent,
having no Ac, introduced a chromosome 9 with C and
Ds (standard location), The majority of the kernels
are similarly variegated for sectors of the c genotype
due to breaks at Ds that occurred in the C Ds chro-
mosome during the development of the kernels. Note
the few fully colored kernels in which ¢ sectors are
absent, and also the several kernels that show large
sectors of the c genotype.

for another generation. For example, if a plant
contains two nonlinked Ac factors, the gametic
ratios approach 1 two-Ac:2 one-Ac:] no-Ac—that
is, a three-to-one ratio for the presence of Ac.
On ears derived from crosses in which such plants -
‘are used as male parents, the kernels with one
or with two doses of Ac may be distinguished be-
cause of clearly seen differences in the time of
response of Ds to Ac doses. Therefore, some of
the kernels considered to have two Ac factors
and others considered to have only one Ac factor
were selected from the test ears. The plants
grown from them were again tested for gametic
ratio of Ac. In each case, verification was
realized. The gametic ratios produced by the
latter plants approached 1 one-Ac:1 no-Ac,
whereas those produced by the former approxi-
mated 3 with one or two Ac factors to 1 with no Ac

The above-described series of tests, and still
others that have been concerned with the time
and type of changes occurring at Ac, have made
it possible to understand the nature of its in-
heritance patterns. It has been found that, with
any particular state or dose of Ac, the time of
occurrence of changes of Ac is controlled by Ac
itself. If, with a particular Ac state, the time
of such changes is delayed until late in the
development of the endosperm, then all the ker-
nels should show this same late timing. This is
known to occur with many of the isolates of Ac.
As the photograph in Figure 23 has shown, how-
ever, a few aberrant kernels may be present on
some of these ears. Some internal or external
alteration in environmental conditions may have
caused these few early-occurring changes at Ac.
No attempts have been made, however, to study
conditions that might alter the time of such
changes.

If these tests for determining the inheritance
behavior of 4c had not been made, considerable
confusion might have arisen. This would certainly
have been true had states of Ac giving relatively
early changes been used in the initial inheritance
studies. It must be stated that just such a situa-
tion has been observed. States of Ac giving aber-
rant gametic ratios have arisen. It is now real-
ized that this is to be anticipated. It has been
determined that the reason for the difference in
patterns of inheritance between an Ac isolate
that gives clear-cut mendelian gametic ratios
and one of its modified derivatives, that gives
aberrant gametic ratios, is related to the time in
the development of the sporogenous or gameto~
phytic cells at which such changes in Ac arise.
With reference to the gametic constitutions that
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 41

will be produced, the time when these changes
occur is most critical. If they occur in somatic
divisions before the meiotic mitoses, or in the
male or female gametophytes, an apparently un-
orthodox inheritance pattern for Ac will result.
If they occur late, that is, in the endosperm
tissues—which act in this connection like a con-
tinuation of the development of the gametophyte—
then no such confused pattern of inheritance will
arise. The gametic constitutions will then closely
approximate those predicted for mendelizing units.
In the study of Ac inheritance, it was necessary
to make selections for these latter states of Ac.
A few exceptions with regard to the time of
changes at Ac may occur in some cells, even

with such selected states of Ac. It was the anal- |

ysis of Ac constitutions in plants derived, in
cell lineage, from those cells in which such ex-
ceptional timing of changes at Ac had occurred,
that provided the information leading to apprecia-
tion of the somatic origins of altered states and
locations of Ac.

Confused patterns of inheritance behavior of
mutable loci have been described in the literature
many times. The ratios obtained have often been
so irregular that no satisfactory formulation of
the nature of the inheritance patterns could be
derived. This would be just as true of some of
the autonomous mutable loci in maize if attention
had not been given to altered states and their
behavior. Two examples may illustrate this. Both
a,™-1 and a,™-1, when first discovered, produced
many mutations and changes in state very early
in the development of the plant. The plant, there-
fore, was sectorial for the altered conditions at
these mutable loci. The sectors were present
in the tassel. When pollen was collected without
reference to the sectors present, and placed on
the silks of plants carrying the stable a, or a,
alleles, the kinds of kernels appearing on the
resulting ears, and their frequencies, were not
readily analyzable in terms of mendelian ratios.
No such difficulty arises, however, when similar
tests are made for gametic ratios in plants de-
rived from those kernels on the original test ears
that show only very late-occurring mutations. The
inheritance pattern is now of the obvious mendelian
type, for mutations and changes in state are
mainly delayed until after meiosis and gamete
formation. As with Ac, the selection of states
of autonomous mutable loci that produce very
late-occurring mutations makes it possible to
examine the inheritance behavior of such loci,
freed from the apparent confusion resulting from
early-occurring modifications at the locus, which
can distort the expected mendelian ratios.

Is THE BEHAVIOR OF Ac A REFLECTION OF A
MECHANISM OF DIFFERENTIATION?

We now return to the original question. What
is the significance of the somatically occurring
changes at Ac, and the changes in state that
occur at the autonomous mutable loci? Do they
suggest the presence of nuclear factors that
serve to control when and where certain decisive
events will occur in the nucleus? With regard
to Ac, it is known that the events leading to its
loss, to increase or decrease of its dosage action,
or to other changes involving its action or position
in the chromosome complement, are related; and
that they appear as the consequence’ of a mitotic
event, controlled in time of occurrence by the
state and dose of Ac itself. Sister nuclei are
formed that differ with respect to Ac constitution,
as the photographs of Figures 17 to 22 illustrate.
Because of this somatically occurring event in-
volving Ac, the Ac-controlled mutable loci will
differ markedly as to the time when mutational
events will occur at them, or as to whether or
not any such events will occur at all in the cells
arising from the sister cells. This precise tim-
ing of somatic segregations effects a form of
differentiation, for it brings about changes in
the control of occurrence and time of occurrence

’ of genic action at other loci, and does so differ-

entially in the progeny of two sister cells. This
likewise applies to the autonomous mutable loci;
but in these cases the controller of the time and
place of appearance of genic activity is a com-
ponent of the locus itself,

The process of differentiation is basically one
involving patterns of action arising in sequential
steps during development and affecting the types
of activities of definitive cells. The ultimate
expression of component parts of an organism
represents the consequence of segregation mech-
anisms involving the various cellular components.
The part played by any one component of the cell
in this segregation system can not be divorced
from that played by any other component. It is
possible, however, to attempt to examine the
various components in order to determine their
respective relationships and the sequential
events that involve them. Embryological studies
have contributed much to our knowledge of the
segregation of cytoplasmic components. The seg-
regation of nuclear components is less well under-
stood, although some outstanding examples are
known. These examples show segregations or
losses of obvious components of the nucl eus—that
is, of whole chromosomes or easily seen parts of
chromosomes. The segregation or loss of smaller
components, not readily visible on microscopic
42 BARBARA McCLINTOCK

examination, may well be one of the mechanisms
responsible for the nuclear aspects of control of
the differentiation process. The phenomenon of
variegation, as described here and observed in
many other organisms, may be areflection of such
a segregation mechanism—exposed to view be-
the timing of events leading to a specific type of
genic action is ‘‘out-of-phase”’ in the developmen-
tal path. Variegation may represent merely an ex-
ample of the usual process of differentiation that
takes place at an abnormal time in development.
Viewed in this way, it is possible to formulate
an interpretation of the’ part played by the nu-
clear components in controlling the course of
differentiation.

—'This interpretation considers that the nucleus
is organized into definite units of action, and that
the potentials for types of genic action in any one
kind of cell differ from the potentials in another
kind of cell. In other words, the functional capa-
cities of the nuclei in different tissues or in dif-
ferent cells of a tissue are not alike. The dif-
ferences are expressions of nonequivalence of
nuclear components. This nonequivalence arises
from events that occur during mitotic cycles.
The differential mitotic segregations are of sev-

___eral types. Some involve controlling components,

such as Ac, and produce sister nuclei that are no
longer alike with respect to these components.
As a result, the progeny of two such sister cells
are not alike with respect to the types of genic
action that will occur. Differential mitoses also
produce the alterations that allow particular genes
to be reactive. Other genes, although present,
may remain inactive. This inactivity or suppres-
sion is considered to occur because the genes
are “‘covered” by other nongenic chromatin mate-
rials. Genie activity may be possible only when
a physical change in this covering material al-
lows the reactive components of the gene to be
“texposed’’ and thus capable of functioning.

A mechanism of differentiation that requires
differences in nuclear composition in the various
cells of an organism finds considerable support
in the literature. The most conspicuous example
is in Sciara, where a thorough cytological and
genetical analysis has been made. (For reviews,
see Metz, 1938, and Berry, 1941.) It is known,
in this organism, at just what stage of develop-
ment differences in nuclear composition will
arise; and, with regard to the X chromosome, it
is known what element in the chromosome con-
trols the differential behavior. This element is
at or near the centromere of the chromosome
(Crouse, 1943), Furthermore, differential segre-

gations of the B-type or accessory chromosome
have been found to occur in a number of plants.
(For reviews of literature to 1949, see Miintzing,
1949.) Numerous other examples are known of
differential segregation involving whole chromo-
some complements, certain types of chromosomes
of a complement, or, occasionally, a certain com-
ponent of a chromosome. (For literature citations
see Melander, 1950; White, 1945, 1950; Berry,
1941.) Whether the differential segregations in-
volve whole complements of chromosomes, indi-
vidual chromosomes, individual parts of chromo-
somes that can be seen, or submicroscopic parts
of chromosomes, may well be a matter of degree
rather than type. Certainly, the evidence for
differential segregation is not wholly negative.
With regard to mechanisms associated with dif-
ferentiation and genic action, an additional factor
may be mentioned. The part played by the doses
of component elements in the chromosomes appears
to be of considerable importance. First, a num-
ber of genetic factors associated with known loci
produce measurable quantitative effects that are
related to dose: the higher the dose the greater
the effect. Such dosage actions, probably reflect
ing rates of reaction, are familiar to all geneti-
cists, and some of them have been reviewed in
this study of mutable loci. Dosage controls of
the Ac type, affecting the time of action of cer-
tain other factors carried by the chromosomes,
has been less well appreciated. A third type of
dosage ‘action has made itself evident in these
studies. In some aspects, however, it resembles
the action of different doses of Ac. In the study
of the autonomous 4,”°! mutable locus, a number
of mutants appeared, particularly on self-pollinated
ears, showing a pale aleurone color. Study of
the behavior of these pale mutants has revealed
the following. Some of them produce pale-colored
aleurone in one, two, and three doses and give no
evidence of instability in the expression of the
phenotype. (One and two doses are obtained by
combinations of the pale-mutant allele with the
stable a, allele.) That this stable expression
may be deceptive is shown by the dosage effects
of other similarly appearing pale-producing iso-
lates derived from mutations at a,"-1. These may
give pale aleurone color, and no indication of in-
stability, in three and two doses; but with one
dose something unexpected occurs. The kernels
show a colorless aleurone in which mutations to
deep aleurone color appear. Still other isolates
give pale color in three doses, but in one or two
doses produce the colorless background with
deep-colored mutant areas. In these cases, it
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 43

is clear that some of the mutations at a,"*! giving
pale color and appearing to be stable are stable
only because of some dosage action produced by
a mutation of the original a,”-! to the pale-
producing type.

The study of dose-provoked actions in the pale
mutants mentioned above and those of Ac have
given some indication of the importance of dosage
action in affecting genic expression. The original
isolate of a,7-1 did not give evidence of such
striking dosage action. When present in one, two,
or three doses, it gave rise to colorless kernels
in which mutations, mainly to a deep color and
occasionally to a pale color, appeared. The
graded series of dosage action exhibited by the
various pale mutants derived from a,”-! is very
much the same as the graded series exhibited by
the various isolates of Ac. In these cases, it
appears as if each isolate is composed of a spe-
cific number of reactive subunits and that the
dosage expressions are related to the total num-
ber of such units that are present in the nucleus.
Although these graded dosage effects may be
visualized on a numerical basis, it is not claimed
that such an interpretation is necessarily the
correct one. The large differences in dosage
expression exhibited by the various isolates of
Ac, and also the various isolates of the pale
mutants derived from a,”°!, nevertheless appear
to follow such a scheme.

Why different doses of components of the chro-
mosomes function as they do in controlling de-
velopmental processes takes us to another level
of analysis that is not under consideration here.
A relation to rates of particular reactions can be
suspected. It is tempting to consider that changed
environmental conditions may well alter otherwise-
established rates of reaction, and thus initiate
alterations in the nuclear components at predict-
able times, leading to strikingly modified pheno-
typic expression. Just such effects have been
observed by students of developmental genetics.
They have shown that alterations of environmental
conditions at particular times in development can

lead to predictable changes in the subsequent
paths of differentiation.

CONSIDERATION OF THE CHROMOSOME
ELEMENTS RESPONSIBLE FOR
INITIATING INSTABILITY

It will be recalled that this study of the origin
and behavior of mutable loci was undertaken be-
cause a large number of newly arisen mutable
oc appeared in the progeny of plants in which an
unusual sequence of chromosomal events had

occurred—that is, the breakage-fusion-bridge
cycle. Striking similarities in the patterns of
behavior of these mutable loci were immediately
noticed. It was the pattern of behavior, rather
than the change in expression of the particular
phenotypic character, that was obviously of im-
portance. This pattern, revealed in all cases,
stemmed from an event occurring at mitosis, which
altered the time and frequency of mutations that
would subsequently occur in the cells derived
from those in which this event occurred. It was
noticed. that sister nuclei could differ in these
respects—and sometimes reciprocally, as if the
mitotic event had resulted in an increase in one
nucleus of a component controlling the mutation
time or frequency, and a decrease of this com-
ponent in the sister nucleus. It was also noted
that the change in phenotypic expression—that
is, the mutation—likewise resulted from a mitotic
event; and that the mutation itself and changes
of the controller of the mutation process could
result from the same mitotic event: one cell show-
ing the mutation, the sister cell showing an al-
tered condition with respect to control of future
mutations in the cells derived from it.

Further, it may be recalled that the mechanism
which resulted in the appearance of newly arisen
mutable loci~that is, the breakage-fusion-bridge
cycle involving chromosome 9%—gave rise to
numerous obvious alterations of the heterochro-
matic materials, in other chromosomes of the com-
plement as well as in chromosome 9. It was also
demonstrated that the effect of a known activator,
Dt, located in the heterochromatin of the chromo-
some-9 short arm, and producing a very definite
pattern of mutations of the otherwise stable a,
locus in chromosome 3, could be recreated inde-
pendently and on a number of different occasions
in cells of a tissue in which the breakage~fusion-
bridge cycle was in action. The combined ob-
servations and experiments point to elements in
the heterochromatin as being the ones concerned
with differential control of the times at which
certain genes may become reactive. It is believed
that somatic segregations of components of these
elements may initiate the process of nuclear con-
trol of differentiation.

On the basis of these interpretations and those
given in the previous section, it becomes apparent
why a large number of newly arisen mutable loci
appeared in the self-pollinated progeny of plants
that had undergone the chromosome type of
breakage-fusion-bridge cycle. This cycle induced
alterations in the heterochromatin, These altera-
tions changed the organization of the heterochro-
44 BARBARA McCLINTOCK

matic chromosome constituents and probably also,
in many cases, the doses of their component ele-
ments. Changes were induced in these hetero-
chromatic elements at times other than those at
which they would normally occur during differ-
entiation. This resulted in changes in the times
in development when their action on specific
chromatin material, associated with genic com-
ponents of the chromosome, was expressed. The
altered timing of their actions was consequently
“out-of-phase” with respect to the timing that
occurs during normal differentiation. This was
made evident by the appearance of a ‘mutable
locus.”? The “‘mutable locus” is thus a conse-
quence of the alteration of an element of the
heterochromatin produced by the breakage-fusion-
bridge cycle. Once such an “out-of-phase”
condition arises, others may subsequently appear
because of the physical changes in the chromatin
that occur at the mutable locus, leading, at times,
to transpositions of this chromatin to new loca-
tions, as described earlier. In their new locations,
these transposed chromatin elements continue
their specific control of types of genic action but
now affect the action of the genic components at
the new locations.

RELATION OF “MUTABLE Locr’ TO
“PosITION-EFFECT’’ EXPRESSIONS
In DROSOPHILA AND OENOTHERA

In a previous publication (McClintock, 1950)
the author has suggested that the position-effect
variegations in Drosophila melanogaster and the
variegations observed in many other. organisms,
including those associated with the mutable loci
here described, are essentially the same. An
adequate discussion of the interrelations would
require more space than can appropriately be
given here. Attention will be drawn, therefore,
only to a few relevant facts, which may serve to
indicate why this conclusion has been reached.
In the first place, a number of different types of
position-effect expression are found in Drosophila
(for review and literature citations, see Lewis,
1950). In maize, comparable types of instability
expression have appeared. In Drosophila, some
of the variegations appear to result from loss of
segments of chromosomes. This applies to those
cases where the expression of the dominant
markers, carried by the chromosome showing the
“‘position-effect”” phenomenon, is absent in some
sectors of the organism. The extent of the de-
ficiency varies, but it includes in each case the
region adjacent to the heterochromatic segment
with which many of the variegation types of

position-effect expression are known to be asso-
ciated. It may be recalled that such deficiencies
are produced in maize when Ds is present.

That heterochromatic elements of the chromo-
somes of Drosophila undergo breakage events in
somatic cells is suggested by the study of
“‘somatic crossing over’’ in this organism (Stern,
1936). The appearance of the abnormally timed
exchanges between chromosomes is conditioned
by the presence of certain Minute factors, for
example, M(1)n, much as the occurrence of struc-
tural aberrations at certain loci in maize (i.e.,
Ds, wherever it may be) is dependent on the pres-
ence of Ac.

Of particular significance for comparative pur
poses is the study of Griffen and Stone (1941)
on the induction of changes in the position-effect
expression in Drosophila of the white-eye varie-
gation, w5, The wm) case arose through an X-
ray-induced translocation of the segment of the
left end of the X chromosome at 3C2 (the u* locus)
to the heterochromatic region of chromosome 4.
Males carrying w™? were X-rayed, and the progeny
examined for changes in the variegation expres-.
sion of the eye mottling. Many such changes were
found. Studies of these cases were continued in
order to determine the nature of the events asso~-
ciated with the changes. In all cases, the new
modification in the phenotypic expression of the
w* locus was found to be associated with a trans-
location, which placed the segment of the left
end of the X chromosome, from 3C2 to the end of
the arm, at a new location. In many cases, the
new position was to a euchromatic region of
another chromosome, and yet variegation per-
sisted. Some of the new positions, however,
gave rise to apparent “‘seversions” to a wild-
type expression. Individuals having these “‘re-
versions’? were X-rayed, and variegation types
again appeared in the progeny. Here also the

‘variegation was shown to be associated with a

translocation involving the left end of the X
chromosome at 3C2, from the location in the
‘Seversion’? stock to a new location—again,
sometimes a euchromatic region. It may be sus-
pected that the maintenance of variegation poten-
tialities in all these cases was associated with
the presence of a segment of heterochromatin of
chromosome 4 that remained adjacent to the w*
locus when the successive translocations occur-
red. This would not readily be detected in the
salivary chromosomes. “The presence of such
‘“inserted’’ heterochromatin could be responsible
for the continued expression of variegation at the
w* locus in repeated translocations. If such was
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 45

the case, then the resemblance to the maize
cases, described in this report, is obvious. The
appearance of ‘‘reversions,’’ and the subsequent
appearance of variegation after X-radiation of
individuals carrying such “‘reversions,’’ might
seem to present a contradiction, On the basis
of an analysis of the cases described by Griffen
and Stone, the writer believes that no contra-
diction is involved. This analysis has suggested
that the timing of variegation-producing events
during development is, in part, a function of the
relative distance of the translocated segment—
ive., the left end of the X chromosome—from the
centromere of the chromosome that carries it: the
farther removed the segment is from the centromere,
the later in the development the variegation-
producing events will occur. In the “reversions,”
this segment has been placed close to the end
of one arm of a chromosome. The reappearance
of variegation occurs when the segment is trans-
located to a position closer to a centromere.
Another factor is also associated with the timing
of the variegation-producing events. This is
the Y chromosome. When the Y is absent, the
areas of altered phenotype are larger than when
it is present, indicating an earlier timing of the
variegation-producing events. It may be noted
in this connection that some of the cases of
‘reversions’? are only apparent reversions. In
XY constitutions they appear to give a stable
wild-type expression but in XO constitutions,
the eyes show a light speckling of the altered
phenotype. With the latter constitution variega-
tion occurs, but only very late in the develop-
ment of the eye. The similarity of this effect
of the Y chromosome to that of dosage action of
Ac is apparent in these cases as well as in many
others in Drosophila that have been examined.

In a recent report, Hinton and Goodsmith (1950)
gave an analysis of induced changes at the bw?
(Dominant brown eye) locus in chromosome 2 of
Drosophila, This case was considered to be a
stable-type position effect. It arose originally
through the insertion of an extra band next to the
salivary-chromosome band where bw* is located.
Males carrying bwP were irradiated and crossed
to wild-type females, The offspring (9,757 in-
dividuals) were examined for changes in the bw?
expression, Twenty-one individuals showing
the wild-type expression appeared in the F,, and
progeny was obtained from one-third of them. A
study of the inheritance behavior of each modi-
fication was undertaken, and a study was also
made of the salivary-gland chromosomes. From
these studies it was clear that the modifications

arose from changes that occurred in the vicinity
of the bw? locus and involved the inserted band.
In four cases, restoration of the wild-type ex-
pression followed removal of this band. In two
cases, it followed separation of this band from
the bw* band by translocations. In one case, no
obvious change in the salivary chromosomes was
noted, but nevertheless a change in phenotypic
expression had occurred. Of considerable im-
portance, also, was the appearance, in some of
these cases, of somatic instability of expression
of the bwt phenotype. Variegation began to
appear. It had never been observed in the brown-
Dominant stock itself, It may be noted that the
changes in the bw? expression are associated
with types of chromosomal alterations which are
much the same as those proposed to account for
changes in phenotypic expression at some muta-
ble loci in maize.

A further resemblance between Drosophila and
maize will be mentioned. In Drosophila, many
of the translocations, inversions, and duplications
are believed to be associated with the formation
of dominant-lethal effects. In maize, a number
of dominant lethals have arisen from transposi-
tions of Ds. Some produce defective growth of
the endosperm and embryo; others affect the
development of the embryo but not the endosperm;
and still others affect the capacity of the embryo
to germinate, without affecting its morphological
characters. Over half the newly arisen trans-
positions of Ds that are of this latter type have
not produced viable plants, owing to lack of ger
mination of the embryos in the kernels.

There are similarities between the maize cases
and a case in Drosophila pseudoobscura des-
cribed by Mampell (1943, 1945, 1946). In this
Drosophila case, a heterochromatic element
appeared to be associated with the initiation of
instability at another locus, which in turn led
to changes in chromosome organization and to
numerous changes in genic action at various
loci in the chromosome complement. These
changes were expressed both somatically and
germinally.

The position-effect behavior reported in Oeno-
thera (Catcheside, 1939, 1947a, b) is much like
that of Ds. The chromosomal events responsible
for the observed types of change in phenotypic
expression may be the same in the two organisms.
In Oenothera as in maize, gross changes in chro-
mosome constitution arise, such as duplications
and deficiencies of segments of the chromosome
involved. Similar cases in other organisms un-
doubtedly exist. It is probable, however, that
46 BARBARA McCLINTOCK

the lack of a critical mode of detection of a chro-
mosome breakage mechanism has been responsible
for the apparent delay in reporting such cases in
connection with studies of somatic variegation
and mutable loci. Also, because changes in state
occur that involve reduction in the frequency of
chromosome breaks, and because such breaks lead
to lethal gametes, it is probable that states of a
mutable locus producing some detectable breaks
are rapidly eliminated from a population, leaving
a state of the mutable locus that produces few or
no such events to be propagated.

It has been argued that the variegation types
of position effect in Drosophila usually do not give
rise to germinal mutations, and that they belong,
therefore, to a separate category of instability
expression. Since some variegation position
effects do give rise to germinal changes, this
argument could in any case be only partially ap-
plicable. However, whether or not germinal changes
arise is not considered relevant in the interpreta-
tion developed here. The time and place of oc-
currence of such changes is related to controls,
existing in the nucleus. The differentiation mech-
anism described above should effect controls that
would exclude the germ lines from undergoing
many changes, but should allow numerous altera-

tions in the soma that would lead to altered pat-

terns of genic expression. Whether or not a par-
ticular somatically expressed pattern of genic
actionfor example, the distribution of pigment
arises from mutations at a “‘mutable locus’’ or
from the action of a particular ‘‘stable”’ allele of
the locus cannot be decided by using the criterion
of presence or absence of germinal mutations.
The important consideration is when, where, and
how the patterns of genic action are controlled
and eventually expressed.

The combined evidence from many sources sug-
- gests that one should look first to the conspicuous
heterochromatic elements in the chromosomes in
search of the controlling systems associated with
initiation of differential genic action in the various
cells of an organism; and secondarily to other such
elements, which are believed to be present along
the chromosomes and to be either initially or sub-
sequently involved in the events leading to differ
ential genic action. Evidence, derived from
Drosophila experimentation, of the influences of
various known modifiers on expression of pheno-
typic characters has led Goldschmidt (1949, 1951)
to conclusions that are essentially similar to
those given here.

The conclusions and speculations on nuclear,
chromosomal, and genic organization and behavior

included in this report are an outgrowth of studies
of the instability phenomenon in maize. They are
presented here for whatever value they may have
in giving focus to thoughts regarding the basic
genetic problems concerned with nuclear organiza-
tion and genic functioning. Until these problems
find some adequate solution, our understanding
and our experimental approach to many phenomena
will remain obscured.

REFERENCES
BERRY, R. ©., 1941, Chromosome behavior in the

germ cells and development of the gonads in Sciara
ocellaris. J. Morph. 68: 547-576.
CATCHESIDE, D. G., 1939, A position effect in Oeno-
thera, J. Genet. 38: 345-352.
1947a, The P-locus position effect in Oenothera.
J. Genet. 48: 31-42.
1947b, A duplication and a deficiency in Oenothera,
J. Genet. 48: 99-110.
CROUSE, HELEN V., 1943, Translocations in Sciara;
their bearing on chromosome behavior and sex

determination. Res. Bull. Mo. Agric. Exp. Sta.
379: 1-75.

DEMEREC, M., 1935, Mutable genes. Bot. Rev. 1:
233-248.

EMERSON, R. A., 1921, Genetic evidence of aberrant
chromosome behavior in maize. Amer. J. Bot. 8:
411-424.

FOGEL, S., 1950, A mutable gene at the R locus in
maize. Rec. Genet. Soc. Amer. 19: 105.

GOLDSCHMIDT, R. B., 1949, Heterochromatic heredity.
Hereditas, Suppl. 5: 244~255.

GOLDSCHMIDT, R. B., HANNAH, A., and PITERNICK,
L. K., 1951, The podoptera effect in Drosophila
melanogaster. Univ. Calif. Publ. Zool. 59:
67-294.

GRIFFEN, A. B., and STONE, W. S., 1941, The w>
and its derivatives. Univ. Texas Publ. No. 4032:
190-200.

HINTON, T., and GOODSMITH, W., 1950, An analysis
of phenotypic reversion at the brown locus in
Drosophila. J. Exp. Zool. 114: 103-114.

LAUGHNAN, JOHN R., 1949, The action of allelic
forms of the gene A in maize. II. The relation of
crossing over to mutations of Ab, Proc. Nat. Acad.
Sci. Wash. 35: 167-178.

1950, Maize Genetics Cooperative News Letter 24:
51-52.

1951, Maize Genetics Codperative News Letter 25:
28-29.

LEWIS, E.B.,1950, The phenomenon of position effect.
Advances in Genetics 3: 73-115.

MAMPELL, K., 1943, High mutation frequency in Dro-
sophila pseudoobscura, Race B. Proc. Nat. Acad.
Sci. Wash. 29: 137-144.

1945, Analysis of a mutator. Genetics 30: 496-505.
1946, Genic and non-genic transmission of mutator
activity. Genetics 31: 589-597.

MANGELSDORF, P. C., 1948, Maize Genetics Codpera-
tive News Letter 22: 21.

McCLINTOCK, B., 1941, The stability of broken ends
of chromosomes in Zea Mays. Genetics 26:
234-282.
CHROMOSOME ORGANIZATION AND GENIC EXPRESSION 47

1942, The fusion of broken ends of chromosomes
following nuclear fusion. Proc. Nat. Acad. Sci.
Wash. 28: 458-463.

1950, The origin and behavior of mutable loci in
maize. Proc. Nat. Acad. Sci. Wash. 36: 344-355.

MELANDER, Y., 1950, Accessory chromosomes in
animals, especially in Polycelis tenus, Hereditas
36: 19-38.

METZ, C. W., 1938, Chromosome behavior, inheritance
and sex determination in Sciara, Amer. Nat. 72:

. 485-520. :

MUNTZING, A., 1949, Accessory chromosomes in Secale
and Poa. Proc. Eighth Intern. Congr. Genetics.
(Hereditas, Suppl. Vol.).

NUFFER, M. GERALD, 1951, Maize Genetics Codpera-
tive News Letter 25: 38-39.

RHOADES, M. M., 1936, The effect of varying gene
dosage on aleurone color in maize. J. Genet.
33: 347-354.

1938, Effect of the Dt gene on the mutability of the
a, allele in maize. Genetics 23: 377-395.

1941, The genetic control of mutability in maize.
Cold Spring Harb. Symposium Quant. Biol. 9:
138-144. .

1945a, On the genetic control of mutability in maize.
Proc. Nat. Acad. Sci. Wash. 32: 91~95.

1945b, Maize Genetics Cooperative News Letter 20:
14.

1947, Maize Genetics Cooperative News Letter 21: 3.

1950, Maize Genetics Cooperative News Letter 24:
49. .

RHOADES, M. M., and DEMPSEY, E., 1950, Maize
Genetics Codperative News Letter 24: 50.

SAGER, R., 1951, On the mutability of the waxy locus
in maize. Genetics (in press).

STERN, C., 1936, Somatic crossing over and segrega-
tion in Drosophila melanogaster. Genetics 21:
625-730.

WHITE, M. J. D.,1945, Animal Cytology and Evolution.
Cambridge Univ. Press.

1950, Cytological studies on gall midges (Cecidomyi-
dae). Univ. Texas Pub. No. 5007: 1-80.