MUTATIONS IN MAIZE AND CHROMOSOMAL ABERRATIONS
IN NEUROSPORA

Barsara McCuiintock

During the past year, studies of muta-
tion-controlling systems in maize were
continued at Cold Spring Harbor, and ex-
aminations of chromosomal aberrations in
Neurospora were carried out at the Cali-
fornia Institute of Technology.

Mutations in Maize

In previous Year Books, evidence has
been presented concerning the nature of
action of mutation-controlling systems in
maize that are responsible for the behavior
of mutable genes. Much of the detailed
evidence was derived from study over a
period of years of the system composed of
the two chromosomal units, Ds and Ae.
Extensive examination of other systems
was postponed until the behavior of this
system had been well explored. Knowl-
edge of gene-controlling systems in maize
is progressing rapidly through the efforts
of investigators at other institutions who
are examining some of the gene-controlling
units that have appeared in their materials.
One such unit, being investigated by Dr.
R. A. Brink and his students at the Uni-
versity of Wisconsin, behaves much like
Ac. My studies of the Ds-Ac system dur-
ing the past year have been limited to ob-
taining some additional information about
. its action at one selected locus in the chro-
imosome complement. This has involved
further examination of the seven cases of
change in action of I induced by Ds when
the latter was located just to the left of Shi
in chromosome 9. The origins and general
descriptions of these cases were outlined on
pages 231-233 of Year Book No. 52. The
newly acquired information provides a
clearer understanding of the similarities
and differences among them. Only a sum-

mary of this information need be presented
here.

Among the seven cases in which the ac-
tion of I (located approximately four cross-
over units to the left of Ds in the parent
plants) was modified, Ds has been found
to be present in the chromosome in each
case and apparently not changed in loca-
tion by the event that altered genic action
in the segment of chromatin to the left of
it. Ds remains near, and to the left of, SA1.
The seven cases may be separated into two
main classes: four that regularly produce
viable homozygotes, and three that have
given no viable homozygotes even though
many attempts have been made this past
year to obtain them. In all seven cases, the
chromosome 9g having the altered region
that includes the locus of J is transmitted
normally through the female gametes; but
transmission through the pollen grains is
considerably reduced when these compete
with others carrying normal chromosomes
g. The extent of this reduction is approxi-
mately the same for all the members of the
first class and for two of the members of
the second class, but is extreme for the
remaining member of the second class.
The seedlings derived from the homozy-
gotes produced by members of the first
class have all shown the same type of al-
bescence phenomenon—disappearance of
chlorophyll from the seedling leaves at
about the three-leaf stage. Intercrossing
has produced individuals carrying all pos-
sible combinations of the seven altered
chromosomes 9, except combinations be-
tween those of class 2, which are apparently
as inviable as are the homozygotes in each
member of this class. Kernels with viable
combinations show no pigment in the
aleurone layer, resembling in this respect
DEPARTMENT OF GENETICS

the phenotypes of the homozygotes of
class 1. The seedlings derived from them
exhibit the same type of albescence phe-
nomenon that characterizes the seedlings
of the class-r homozygotes. In none of the
seven cases has there been any evidence of
crossing over within the affected region in
individuals having a normal and an altered
chromosome 9. In each case, however,
crossing over to the right of the affected

region—between SA: and Bz and between _

Bz and Wx—is normal.

The above-described relations among the
seven independent cases of change in genic
action within a segment of chromosome 9
located to the left of Ds are consistent with
a hypothesis that they originated through
removal of a similar segment of chromo-
some g in each case. The common segment
removed would include the region: begin-
ning just to the left of the locus of SA: and
ending somewhere beyond that of I. The
differences between members of the two
classes and among members of the second
class could then be explained by differences
in the extent of loss beyond the locus of
I—the members of class 2 having longer
deficiencies than those of class 1, ‘and one
member of class 2 the longest. No direct
evidence of deficiency in any one of these
seven chromosomes was obtained, how-
ever, when each was compared with a nor-
mal chromosome g at the pachytene stage
of meiosis. Also, it is known that defi-
ciency need not be involved in such loss
of genic action. A Ds-induced inhibition
of action may be the cause, for this is
known to account for some similar types
of loss of genic action in the segment of
chromosome g located to the right of Ds.
The types and extents of such loss were
described in Year Book No. 52.

System controlling genic expression of
ai", The occurrence of genic instability
in many organisms, and its general resem-
blance to cases in maize, suggest that the
gene-controlling units found in maize are
not peculiar to that species. Rather, they
reflect a type of nuclear organization and
action that is probably present in many

255

organisms. Thus it is necessary, before
wider integrations can be made, to deter-
mine the kinds of gene-controlling sys-
tems that exist, how they differ from one
another, and whether or not interactions
occur among them. Attention is now be-
ing focused, therefore, on a system control-
ling genic expression which differs. from
the Ds-Ac system in some respects but
resembles it in others. Although knowl-
edge of the behavior of this second system
is limited, it has been possible to formulate
an interpretation from present informa-
tion. This system operates to control genic
expression of a:”"*, one of the mutable con-
ditions that has arisen at the locus of A:
in chromosome 3 in the Cold Spring Har-
bor cultures: (A: is associated with the
appearance of anthocyanin pigmentation
in plant tissues and in the aleurone layer of
the kernel; a1, known recessive allele, when
homozygous, results in absence of antho-
cyanin pigmentation in both plant and
aleurone.)

The system controlling genic expression
at a" is considered to be composed of
two units, one inserted at the locus of i
in chromosome 3 at the time of origin of
ai", and one located elsewhere in the
complement. The evidence suggests that
this latter unit, tentatively designated as
Spm for reasons that will be made evident
shortly, may be altered in somatic cells
and possibly may also be transposed from
one location to another in the chromosome
complement. In these respects, its behavior
resembles that of Ac. Another apparent
similarity is that Spm must be present in
the nucleus if mutations are to occur at
the locus of a:.%*. Spm and Ac differ in a
significant way, however, and this is re-
lated to the contrasts in phenotype that
are expressed in their presence and ab-
sence. The presence of Ac is detected by
the alterations it induces at the locus of
the gene whose mutations it controls (or
at the locus of Ds), which lead to recog-
nizable phenotypic modifications in so-
matic cells; its absence is indicated by the
absence of any such modifications. With
256

regard to Spm, however, its presence or
absence is recognized not only by the ap-
pearance or nonappearance of mutations at
a: +, but also by another phenotypic differ-
ence that is sharply expressed in plants
having certain states of a:”*. When Spm
is present in somatic nuclei, genic action at
the locus of a"? is completely inhibited
except in those cells that carry a mutation
at this locus. Such mutations occur in
some somatic cells during development.
They permit the occurrence of some forms
of action of the genic material at the locus
of a", and this is recognized pheno-
typically in the tissues produced by the
descendants of such cells. Therefore, both
plants and kernels that have Spm are
variegated in anthocyanin pigmentation.
Spots or areas, each exhibiting a particular
type and intensity of pigmentation—for
different types of mutations are induced at
a:”* when Spm is present—appear on a
colorless background. Removal of Spm,
whether occurring somatically in some cells
or produced as the consequence of meiotic
segregations in plants heterozygous for it,
results in partial release of inhibition of
genic action, presumably at a:"%, for the
tissues of both the plant and the aleurone
layer of the kernel are uniformly pig-
mented. With some of the states of a",
the color is less intense than that which ap-
pears when the normal A; locus is present.
The difference may be qualitative as well
as quantitative, for the pigment does not
seem to be the same as that produced by
A;. No spots or areas with altered pigmen-
tation appear. In other words, no variega-
tion is expressed in the absence of Spm.
The symbol Spm has been used to desig-
nate this unit because it acts both as a sup-
pressor and as a mutator.

The origin of a:”* from a modification
occurring at the locus of the normal A:
gene was described in Year Book No. 50
(1950-1951). Spm was first discovered in
a study of a particular state of a:"*, de-
rived from the original state. This derived
state was selected early in the investigation
of a+ because it was considered to be

CARNEGIE INSTITUTION OF WASHINGTON

more suitable than the original state for
an analysis of factors responsible for muta-
tion at ai”. It was present in an indi-
vidual kernel in the progeny of the plant
in which a:”" first appeared. Some aspects
of the inheritance behavior of this state of
ai" were outlined in Year Book No. 52.
The evidence described there indicated the
presence in variegated plants of a unit
factor (or of several similar unit factors),
not linked to the locus of a”, that in-
fluenced the expression of a:"*. It was
then thought that removal of this factor
from nuclei by meiotic segregations re-
sulted in a particular type of mutation at
the locus of a:"", and that in this respect
the case resembled that of a:%*, where re-
moval of such a factor is known to lead to
mutation at ai". Results of subsequent
investigation, however, carried out during
the past year, suggest that removal of the
unit factor, now designated Spm, does not
lead to mutation at a7, but rather re-
moves an inhibitory action ascribable to
Spm, as outlined above.

Evidence suggesting linkage of an Spm
factor with Y or its allele y (Y, yellow
starch in endosperm, dominant to y, white
starch), located in chromosome 6 of the
complement, was presented in Year Book
No. 52 in table 3 of the report of this
Department. An exploratory test involv-
ing only nine plants was made this year,
partly in order to determine whether or not
the apparent linkage could be validated.
These plants were derived from the varie-
gated kernels produced on the ear of plant
6046C-2, recorded in the above-mentioned
table. Two or more independently lo-
cated Spm units were assumed to be pres-
ent in this plant, but one of them appeared
to be linked to Y. It should be emphasized
that, in backcross tests, segregation ratios
typical of linked units would appear only
on the ears of plants having a single Spm
unit, and in such plants only when losses
of Spm from many cells did not occur late
in the development of the ear. Further-
more, transpositions of Spm to new loca-
tions would distort ratios or even eliminate
DEPARTMENT OF GENETICS

evidence of linkage if they occurred early
in the development of an ear or if they
were frequent during the later stages of
development... An early-occurring loss of
Spm from a cell whose descendants con-
tribute to the formation of the ear is de-
tected readily by the appearance of an area
on the mature ear in which all the kernels
carrying @:”* show the characteristic pale

257

cated approximately 38 crossover units
from Y. The presumed location in chro-
mosome 6 of one of the Spm units in the
parent plant, 6046C-2, has thus been con-
firmed.

Further description of the behavior of
this gene-controlling system will be post-
poned until the evidence from tests now
under way becomes available.

”

TABLE 17

PHENOTYPES OF KERNELS APPEARING ON EARS OF PLANTS WHOSE CONSTITUTIONS WERE a1” Sha/as sha;
Y/y WHEN THESE WERE CROSSED BY PLANTS HOMOZYGOUS FOR 4, Shs, AND y

 

PHENOTYPES OF KERNELS

 

 

 

 

Shs sha
PARENTAGE Color in aleurone Color in aleurone Totals
IN CROSS -

Pale Variegated Colorless Pale Variegated Colorless

g 3 Y y Y y Y oy Y oy Y¥ y Y¥ y
6629A-1 1041-5... 36 58 66 35 0 0 0 0 0 0 85 89 369
6629A-1 1041-4... 33 54 51 36 0 1 0 0 1 1 101 98 376
6629A-3 1040-1... 34 52 43 37 0 0 0 1 1 0 85 83 336
6629A-4 1041-5... 23 65 56 36 0 0 0 0 0 0 90 84 354
6629A-6 1040-1... 34 67 78 37 0 0 0 0 J 0 86 113 416
6629A-7 1040-1... 29 59 58 36 4 1 0 0 0 0 105 100 392
6629A-9 1040-8... 39 41 49 38 0 0 1 0 0 0 79 90 337
6629A-9 1040-5... 41 58 50 34 0 1 0 1 0 0 79 «115 379
Totals ......... 269 454 451 = -289 4 3 1 2 3 1 710 772 2959

 

aleurone color that develops in the absence
of Spm. A few ears had such areas. Early-
occurring transpositions of Spm can be de-
tected by markedly altered linkage rela-
tions between the given units, by absence
of such linkage, or by linkage of Spm to a
factor carried in another chromosome.
There was evidence of such changes in
location of Spm, based on observed ratios
produced by some of the ears.

The conditions, indicated above, that are
required for the production of typical link-
age ratios in backcross tests were present
in eight ears obtained from six of the nine
tested plants. The design of the test cross
and the types of kernels that appeared on
each of these eight ears are shown in table
17. From the data in this table it may be
concluded that a single Spm unit was pres-
ent in these six plants, linked to Y and lo-

CHROMOSOME ABERRATIONS IN
Neurospora

During the winter of 1953-1954, exami-
nation of chromosome complements in asci
of Neurospora crassa was undertaken at
the California Institute of Technology, for
the purpose of determining the nature of
chromosomal aberrations known to be
present in strains 4637 and 45502. All the
stocks used in this study were obtained
from Mary B. Mitchell, and all crosses
were made by her. Her co-operation and
interest were very much appreciated.

Previous investigations had shown that
strain 4637 carries a reciprocal transloca-
tion between two nonhomologous chromo-
somes. Although the earlier studies were
limited, they did indicate that chromo-
somes 1 and 6 are involved in the transloca-
tion. The purpose of this year’s study was
258

to determine the position of break, in each
of these two chromosomes, that gave rise
to the translocation. Crosses were made
between the wild-type strain 2522-1a and
strain 4637R-1A. Nuclei in asci produced
by this cross were examined at that stage
of the ‘first meiotic prophase when the
chromosomes are maximally elongated. It
was possible, by observing synaptic con-
figurations of the chromosomes involved in
this translocation and by comparing chro-
momere morphologies of paired elements
within it, to determine the break position
in each of the chromosomes.

The linear organization of chromosome
6, the next to the smallest chromosome of
the complement, is unusually well defined,
and this makes possible a ready identifica-
tion of its component parts. That of chro-
mosome 1, the longest chromosome of the
complement, is less well defined; but two
conspicuous chromomeres are present,
which can serve as points of orientation in
describing break positions. One of them
is located near the middle of the chromo-
some, dividing it into two segments with
relative lengths of 1 and 1.6. The other
conspicuous chromomere is located near
the end of the shorter of these two seg-
ments. The position of the break in chro-
mosome 1 is in the longer of its two seg-
ments, approximately three-eighths of the
distance from the free end of the segment.
Chromosome 6 may be divided into several
well marked segments. One terminal seg-
ment contains large, closely packed chro-
momeres. This is followed by a longer
segment with smaller, less closely aligned
chromomeres. A very conspicuous, deep-
staining, dumbell-shaped chromomere sep-
arates the second segment from the remain-
ing segment of the chromosome. This
last segment, whose length is about one-
quarter that of the total chromosome, con-
tains a few small, rather widely separated
chromomeres. The position of the break in
chromosome 6 is in the second of the
above-described segments, closer to the
first segment than to the conspicuous
dumbbell-shaped chromomere.

CARNEGIE INSTITUTION OF WASHINGTON

In crosses of strain 4637 to wild type,
various types of synaptic configuration
were produced by association of the two
translocation chromosomes introduced by
strain 4637 and the two corresponding nor-
mal chromosomes introduced by the wild-
type strain. Homologous association of all
components was frequent, resulting in the
formation of a cross-shaped configuration
having three short arms and one long arm.
Failure of association of homologous com-
ponents of one or more of the shorter arms
of the cross was observed, however, in a
number of nuclei. The remaining five
pairs of chromosomes appeared to be nor-
mal, for no gross structural modifications
were observed in any member of a pair.

In contrast to the relative ease with
which the structural modification present
in strain 4637 could be analyzed, that as-
sociated with strain 45502 proved to be
very difficult. In crosses of this strain to
wild type, a distinctive pattern of normal
and abnormal spores appeared in many of
the asci. Cytological examination indicated
that there is no simple reciprocal transloca-
tion between two nonhomologous chromo-
somes in this strain to account for the
distinctive patterns of spore types. The
chromosome complements of four isolates
of ‘strain 45502 were examined: T45502-
P1315-1a, T45502-1507-1A, T45502-1508-1a,
and 70007TR-2A. The examinations sug-
gested that in each of them the comple-
ment is composed of seven haploid chro-
mosomes plus an extra chromosome whose
length is approximately half that of the
smallest chromosome of the complement,
namely, chromosome 7. The origin of
this extra chromosome could not be deter-
mined readily by inferences from synaptic
associations, because of the many irregu-
larities in chromosome associations that oc-
curred when these strains were crossed to
wild type and even when two of them,
T45502-P1315-1a and 70007TR-2A, were in-
tercrossed.

A brief description of some of the ab-
normal synaptic associations that were ob-
served at the first meiotic prophase in asci
DEPARTMENT OF GENETICS

produced by crossing any one of the four
strains of 45502 with wild type will indi-
cate the nature of the complexities encoun-
tered in this analysis. One end of the frag-
ment chromosome was often associated
with an end of another chromosome. This
did not necessarily reflect homology, be-
cause any one of the seven chromosomes
of the haploid complement could enter
into such an association. It occurred most
frequently, however, with chromosome 7.
When one end of the fragment chromo-
some was associated with an end of an-
other chromosome, the homologue of this
second chromosome sometimes formed a
terminal association with a member of still
another pair of chromosomes. Although
most of the aberrant associations occurred
between ends of chromosomes, a few nu-
clei were observed in which longer seg-
ments of two nonhomologous chromo-
somes were synaptically associated.

Because of these irregularities in synap-
tic behavior, determination of chromosome
composition and organization in the ascus
nuclei was difficult. In all nuclei, pairing
occurred between some of the homologues.
The same pairs, however, were not present
in each nucleus. Nevertheless, by noting in
each nucleus which chromosomes were
paired and by comparing the members of
each pair with regard to their internal
structure, it was possible to determine
whether or not any structural rearrange-
ment was present in any member of a
given pair. For two of the four examined
isolates of 45502, this method produced no
evidence of gross structural modifications
in any of the seven regular chromosomes
of the complement. In the other two iso-
lates, the composition of chromosome 7
was not determined with certainty, but the
other six chromosomes appeared to be
normal.

As was stated earlier, the origin of the
small extra chromosome present in each of
the four isolates of 45502 was not deter-
mined. The high frequency of association
of this fragment with a member of the
chromosome 7 pair suggests a possible deri-

259

vation from chromosome 7. Some support
for this inference is given by the similar
high frequencies of such associations that
appeared in many of the nuclei produced
by crossing 45502-P1315-1a with 7oo07TR-
2A. Each parent presumably contributed
the seven haploid chromosomes plus the
fragment. Thus it was to be expected that
nuclei with eight homologously associated
pairs of chromosomes would be numerous.
Instead, they were relatively rare. More
often, one or both of the fragment chromo-
somes were synaptically associated with
another chromosome of the complement.
Sometimes there appeared very complex
configurations, involving members of sev-
eral different chromosomes of the comple-
ment. Chromosome 7 was a component of
most of these aberrant configurations.

The aberrant synaptic associations ap-
pearing at the first meiotic prophase in all
crosses involving these four isolates of
45502 were reflected at the late diakinesis
and metaphase stages. Associations be-
tween nonhomologous chromosomes were
noted. Also, univalents were present in a
number of nuclei. No concerted effort was
made, however, to analyze these stages, or
still later ones in the meiotic process.
Therefore it is not known whether or not
many abnormal disjunctions of chromo-
somes occurred at anaphase I as a conse-
quence of abnormal pairing or lack of pair-
ing. Neither is it known to what extent
such disjunctions could contribute to ab-
normal spore formation.

The difficulties encountered in the anal-
ysis of the chromosome constitution of
45502 were intensified in the first period of
observation by the presence of another
chromosomal abnormality, which was later
found to have been introduced by the wild-
type parent. The initial observations were
made in asci produced by the cross of
45502-P1315-1a to wild-type strain 2292-2A.
In addition to the fragment chromosome,
a structural abnormality was noted in one
of the two chromosomes 5. This chromo-
some was considerably longer than its
normal homologue, its length being com-
260

parable to that of chromosome 3. A seg-
ment of uncertain origin, added to or in-
serted into one arm of the chromosome,
was responsible for the increased length.
Synaptic associations were very frequent
between this abnormal segment and the
fragment chromosome. In addition, other
complex types of synaptic configurations
were formed, incorporating not only the
fragment chromosome and the abnormal
chromosome 5, but also one or more other
chromosomes of the complement. Fre-
quently, homologues of chromosome 7 or
of both chromosome 7 and chromosome 2
were components of these configurations.
Only after an analysis had been made of
the chromosomes in asci derived from
other crosses was it realized that the ab-
normal chromosome 5 had been intro-
duced by the wild-type parent 2292-2A.
Confusion in interpreting the chromo-
some constitution of 45502—arising from
this initial lack of knowledge of the consti-
tution of the wild-type parent, 2292-2A,
used in some of the crosses—seriously ham-

CARNEGIE INSTITUTION OF WASHINGTON

pered progress toward a solution of the
main problem, which was to determine the
chromosome abnormality responsible for
the distinctive spore pattern and for the
accompanying false linkages of certain
genetic markers. Nevertheless, the knowl-
edge gained from this experience is of
some general significance. It indicates the
necessity for determining the chromosome
constitution of wild-type and tester stocks
before they are used in crosses requiring
cytogenetic analysis. It also suggests that
some of the discordant results derived from
genetic analyses may be due to undetected
structural modifications of chromosomes
in strains presumed to be normal in their
chromosome constitutions.

Although the study of chromosome con-
stitutions and behavior in strains derived
from 45502 had to be terminated before
the cytogenetic relations had been clarified,
the increased knowledge of chromosome
organization and behavior in Neurospora
which it provided will serve as a guide in
future studies of this type.

BIBLIOGRAPHY

BernsTEIn, M. H. Deoxyribonucleoproteins of
cell nuclei: Sensitivity to X-rays. Nature,
vol. 174, p. 463 (1954).

Berriz, A.M. M. The effects of temperature on
ultraviolet-induced mutability in Escherichia
coli. Proc. Nat. Acad. Sci. vol. 39, pp.
1125-1133 (1953).

Demerec, M. (editor). Advances in Genetics,
vol. 6. 488 pp. New York, Academic Press
(1954).

What makes genes mutate? Proc. Amer.
Philos. Soc., vol. 98, pp. 318-322 (1954).
Hersnzy, A.D. Functional differentiation with-
in particles of bacteriophage T2. Cold
Spring Harbor Symp. Quant. Biol., vol. 18,

PP. 135-139 (1953).

—— Nucleic acid economy in bacteria in-
fected with bacteriophage Tz. . II. Phage
precursor nucleic acid. Jour. Gen. Physiol.,
vol. 37, pp. 1-23 (1953).

Some central problems of viral growth.
In The dynamics of virus and rickettsial in-
fections, pp. 13-15. New York, Blakiston
Co. (1954).

Kaurmann, B. N. See Paicen, K.

Kaurmann, B. P. Chromosome aberrations in-
duced in animal cells by ionizing radiations.

 

 

 

In Radiation biology, ed. A. Hollaender, pp.
627-711. New York, McGraw-Hill Book
Co. (1954).

See McDonatrp, M. R.

Kunirz, M., and M. R. McDonaxp, Ribonu-
clease. Zn Biochemical preparations, vol. 3,
ed. E. E. Snell, pp. g-19. New York, John
Wiley & Sons (1953).

Lasrum, E.L. The effect of generation time on
the delayed appearance of induced mutants
in Escherichia coli. Proc. Nat. Acad. Sci.,
vol. 39, pp. 1221-1227 (1953).

LevinTHAL, C., and N. Visconti. Growth and
recombination in bacterial viruses. Genetics,
vol. 38, pp. 500-511 (1953).

McCuintock, B. Induction of instability at se-
lected loci in maize. Genetics, vol. 38, pp.
579-599 (1953).

McDonatp, M. R. The effects of X rays on di-
lute solutions of crystalline trypsin: Contin-
ued inactivation after termination of irradia-
tion. Brit. Jour. Radiol., vol. 27, pp. 62-63
(1954).

The inactivation of dilute solutions of

crystalline trypsin by X-radiation. I. Kinet-

ics and characteristics. Jour. Gen. Physiol,

vol. 38, pp. 93-103 (1954).

 
DEPARTMENT OF GENETICS

 

and B. P. Kaurmann. The degradation

by ribonuclease of substrates other than ribo-

nucleic acid. Jour. Histochem. and Cyto-

chem., vol. 2, pp. 387-394 (1954).

See Kunitz, M.

Paicen, K. The occurrence of several biochem-
ically distinct types of mitochondria in rat
liver. Jour. Biol. Chem., vol. 206, pp. 9457
957 (1954).

 

261

 

and B. N. Kaurmann. Effects of X-ir-
radiation on amount and composition of
nucleic acids in liver. Jour. Cell. and Comp.
Physiol., vol. 42, pp. 163-178 (1953).

Visconti, N. See LevintHat, C.

von Borster, R. C. See Wuitine, A. R.

Wuitine, A. R., and R. C. von Borster. Dom-
inant lethal and inactivation effects of nitro-
gen mustard on Habrobracon sperm. Ge-
netics, vol. 39, pp. 317-325 (1954).