CYTOLOGICAL INVESTIGATION OF DEVELOPMENTAL
PROGRESS TOWARD DIPLOIDIZATION IN LATE
GENERATION AM~dPHIDIPLOIDS FROM THE CROSS
Phasedsus lunatus L. x P. polystachyus ( L.) B.S.P.
} JIMMY EDWARD WYATT
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The writer wishes to express his gratitude to Dr. A. P. Lorz,
Chairman of his Supervisory Committee, for the patient and understand-
ing guidance given throughout the various phases of this study.
Appreciation is extended to Doctors D. S, A~nthony, VI. F. Nettles,
R. C. Smith, and 8. D. Thom~pson Fo their constructive criticism and
invaluable assistance in the preparation of this manuscript.
The wr-iter is also indebted to h~is w~ife, Joyce, for her devotion,
encouragement, and help while thris worke was in progress, and also for
the help rendered during the preparation of this dissertation.
TABLE OF CONTENTS
ACKNOWJLE D GME NT S ,...............
LIST OF TABLES . . . . . . . . . .
LIST OF FIGURES . .. . . . . .. . .
ABSTRACT . . . . . . . . . . .
INTRODUCTION . . . . . . . . . .
LITERATURE REVIEW ,,,,,,,..........
Natural Amphidiploids ..........
Experimental Amphidiploids .......
Fertility Changes in Auto- and
Interspecific Hybridization in Phaseolus
MATERIALS AND METHODS ...............
Root Tip Sectioning ...........
Flower Bud Sectioning ..........
Pollen Grain Study ...........
Anther Squashes .............
RESULTS AND DISCUSSION ...............
SUMMARY . . . . . . . . . . .
APPENDIX . . . . . . . . . . .
LITERATURE CITED .. ... .. .. . ... .
BIOGRAPHIICA L SKETCH . . . . . . . .
LIST OF TABLES
1. Characteristics of the diploid and amphi-
diploid lima bean plants grown in the
greenhouse and in the field, 1970 . .. ... 38
2. Bud measurements of diploid 'Fordhook'
and amphidiploid lima bean lines and
the melotic stage found in anthers of
each bud .. .. .. .. ... .. .. .. . 45
3. Number of normal and abnormal meiotic
figures from amphidiploid lima bean
plants after several generations of
selfing and selection .. ... .. .. .. 58
4, Number of aborted pollen grains in 100
pollen grains counted from each of 5
flowers from diploid and amphidiploid
plants .. .. .. .. .. ... .. .. .. 83
5. Pollen grain diameters in microns from
5 flowers of plant 3 .. .. .. ... .. .. 85
6. Pollen grain diameters in microns from
5 flowers of plant 5 ... .. ... .. .. 86
7. Analysis of variance of pollen grain
diameters of advanced generations of
amphidiploids derived from the cross
P. lunatus var. 'Fordhook' x _P. po1Y"
stachvus .. .. .. .. ... .. .. .. . 87
8. Pollen grain diameters from 5 flowers
of P. lunatus L. var. 'Fordhook' .. .. .. .. 110
9. Pollen grain diameters in microns
from 5 flowers of plant 2 ,.. .. .. .. . Ill
10. Pollen grain diameters in microns
from 5 flowers of plant 4 .. .. .. .. .. 112
11. Pollen grain diameters in microns
from 5 flowers of plant 6 . .. ... .. .. 113
12. Pollen grain diameters in microns
from 5 flowers of plant 9 .. .. .. .. .. 114
13. Pollen grain diameters in microns
from 5 Flowers of' plant 10 .. .. .. ... 115
14. Pollen grain diameters in microns
from 5 flowers of plant 11 .. .. .. ... 116
15. Pollen grain diameters in microns
from 5 flowers of plant 12 .. .. ... .. 117
16. Pollen grain diameters in microns
from 5 flowers of plant 660 . . . .. 118
17. Pollen grain diameters in microns
from 5 flowers of plant 986 .. . . .. l9
18. Pollen grain diameters in microns
from 5 flowers of plant 985 . .. .. .. 120
19. Pollen grain diameters in microns
from 5 flowers of plant LLLP .. .. .. . 121
20. Pollen grain diameters in microns
from 5 flowers of plant LLLP F2-1 . .. ... 122
21. Pollen grain diameters in microns
from 5 flowers of plant LLLP F2-2 . . . . 123
LIST OF FIGURES
1. Petri dish of melted paraffin containing
plastic vial cap with seven straw sections
and buds to be embedded. The thermometer
was used to regulate the temperature of
the hotplate ... ... ... .. .. .. 28
2. Buds packed into soda straw sections
arranged in the plastic vial cap . ... ... 30
3, Paraffin block before removal of excess
paraffin around the straw sections, and
an empty plastic vial cap . .. .. .. . 31
4. Paraffin cylinders formed in soda straw
sections. Left, straw intact; center,
straw slit lengthwise; right, straw
peeled away from the paraffin cylinder .. .. 32
5. Paraffin cylinder attached to a microtome
block. Opposite sides of the cylinder
have been trimmed parallel to each other . ... 33
6. Face view of trimmed paraffin cylinder
showing trimmed parallel sides and
several embedded bean buds .. .. .. ... 34
7. Diploid lima bean P. lunatus L. var.
'Fordhook', spring, 1970 .. .. .. .... 40
8. Bush character of amphidiploid plant
6, spring, 1970 .. ... .. ... .. .. 41
9. Vining character of amphidiploid
plant 9, spring, 1970 .. . ... ... .. 42
10. Flower color of amphidiploid plants
grown in the field, spring, 1970 .. .. .. .. 43
11. Hypogeal character of amphidiploid seed-
lings transplanted into peat pots ,. .. .. 43
12. Above, photomicrograph of a metaphase plate
of a root tip cell of amphidiploid plant 6
(1050X); below, an interpretive camera
lucida drawing (2000X() . ... ... .. .. 49
13. Above, photomicrograph of a metaphase plate
of a root tip cell of P. lunatus L. var.
'Fordhook' (1050X); below, an interpretive
camera lucida drawing (2000X) . .. ... .. So
14. Above, photomicrograph of a pollen mother
cell of amphidiploid plant 660 at diplotene
(1050X); below, an interpretive camera
lucida drawing (2000X) . .. ... .. .. .. 52
15. Above, photomicrograph of a pollen mother
cell of amphidiploid plant 660 at metaphase
11 (1050X); below, an interpretive camera lucida
drawing showing a total of 44 chromosomes in the
two metaphase 11 configuration (2000X) .. .. 53
16. Above, photomicrograph of a pollen mother
cell of amphidiploid plant 2 at metaphase
11 (1050X); below, an interpretive camera
lucida drawing (2000X) .. .. ... .. .. 54
17. Above, photomicrograph of a-pollen mother
cell of amphidiploid plant 660 at metaphase
II (1050X); below, an interpretive camera
lucida drawing (2000X) . ... .. .. .. .. 55
18. Above, photomicrograph of a pollen mother
cell of plant 3 showing laggards at metaphase
S(1050X); below, an interpretive camera
lucida drawing (2000X) .. ... .. .. .. 61
19. Above, photomicrograph of a pollen mother
cell of plant 2 showing laggards at
telophase I (1050X); below, an interpretive
camera lucida drawing (2000X) . ... .. .. 62
20. Above, photomicrograph of a pollen mother
cell of plant 660 showing laggards at
metaphase 11 (1050X); below, an interpretive
camera lucida drawing (2000X) .. .. ... 63
21. Above, ph~otomicrograph of a pollen mother
cell of plant 2 showing laggards at telophase
S(1050X); below, an interpretive camera
lucida drawing (2000X) . ... .. .. .. .. 64
22. Above, photomicrograph of a pollen mother
cell of plant 660 showing laggards at
telophase 1 (1050X); below, an interpretive
camera lucida drawing (2000X) .. .. .. .. 65
23. Above, photomicrograph of a pollen mother
cell of plant 6 showing laggards at
metaphase II (1050X); below, an interpretive
camera lucida drawing (2000X) .. .. .. .. 66
24, Above, photomicrograph of a pollen mother
cell of plant 660 showing laggards at
telophase II (1050X); below an interpretive
camera lucida drawing (2000X) . .. .. .. .. 67
25. Above, photomicrograph of a pollen mother
cell of plant 660 showing a short chromatin
bridge at early anaphase I (1050X); below~,
an interpretive camera lucida drawing (2000X) . 70
26. Above, photomicrograph of a pollen mother
cell of plant 2 showing several chromatin
bridges at anaphase I (1050X); below, an
interpretive camera lucida drawing (2000X) .. 71
27. Above, photomicrograph of a pollen mother
cell of plant 2 showing several chromatin
bridges at anaphase I (1050X); below, an
interpretive camera lucida drawing (2000X) .. 72
28. Above, photomicrograph of a pollen mother
cell of plant 2 showing a single bridge at
late anaphase I (1050X); below, an inter-
pretive camera lucida drawing (2000X) . ... 73
29. Above, photomicrograph of a pollen mother
cell of plant 2 showing a chromatin bridge
and chromosome fragment at telophase
(1050X); below, an interpretive camera
lucida drawing (2000X) .. .. .. .. .. .. 74
30. Above, photomicrograph of a pollen mother
cell of plant 2 showing a chromatin bridge
and two chromosome fragments at telophase
(1050X); below, an interpretive camera lucida
drawing (2000X) .. .. .... .. . ... 75
31. Above, photomicrograph of a pollen mother
cell of plant 2 showing a chromatin bridge
with four chromosome fragments at telophase
(1050X); below, an interpretive camera lucida
drawing (2000X) .. .. .. .. .. .. ... 76
32. Above, photomicrograph of a pollen mother
cell of plant 2 showing two chromiatin bridges
at telophase I (1050X); below, an interpretive
camera lucida drawing (2000X) .. .. .. .. 77
33. Photomicrograph of a tetrad showing four
large pollen grains and one small pollen
grain (500X) .. ... .. .. .. .. .. .. 79
34. Phottomicrograph of a tetrad showing four
large and two small pollen grains (500X) .. .. 80
35. Photomicrograph of a tetrad showing four
large and three small pollen grains (500X) . .. 81
36. Frequency distribution of pollen grain
diameters of amphidiploid plant 3 . .. .. .. 88
37. Frequency distribution of pollen grain
diameters of amphidiploid plant 5 . .. .. .. 89
38. Frequency distribution of pollen grain
diameters of diploid variety 'Fordhook' .. .. 90
39. Frequency distribution of pollen grain
diameters of amphidiploid plant 2 .. .. .. 91
40. Frequency distribution of pollen grain
diameters of amphidiploid plant 4 ,. .. . .. 92
41. Frequency distribution of pollen grain
diameters of amphidiploid plant 6 . .. .. .. 93
42. Frequency distribution of pollen grain
diameters of amphidiploid plant 9 .. .. .. 94
43. Frequency distribution of pollen grain
diameters of amphidiploid plant 10 . .. .. .. 95
44. Frequency distribution of pollen grain
diameters of amphidiploid plant 11 .. .. .. 96
45. Frequency distribution of pollen grain
diamreters of amphidiploid plant 12 .. .. .. 97
46. Frequency distribution of pollen grain
diameters of amphidiploid plant 660 .. . ... 98
47. Frequency distribution of pollen grain
diameters of amphidiploid plant 986 .. .. .. 99
48. Frequency distribution of pollen grain
diameters of amphidiploid plant 985 . .. .. 00
49. Frequency distribution of pollen rain
diameters of plant LLLP ,, ,, .. .. ... 101
50. Frequency distribution of pollen grain
diameters of plant LLLP F2-1 , 102
51. Frequency distribution of pollen grain
diameters of plant LLLP F2-2 ... .. . 103
Abstract of Dissertation Presented to the
Graduate Council of the University o-f Florida in Partial Fulfillment
of the Requirements for- the Degree of Doctor of Philosophy
CYTOLOGICAL INVE-STIGATION OF DEVELOPMENTAL PROGRESS
TOWARnD DIPLOIDIZATION IN LATE GENdERATIONJ
AM4PHIDIPLi0DS FROM4 THE CROSS
Phascolus lunatucs. L. x _P. Epolystacgus (L.) B.,SP.
J imm~y Ed~ward W~yatt
Chai rman: Or, A. P, Lorz
Maj or Depar-tmnent : Vegetable Crops
The cytoyenetic behavior of an advanced genleraotion, approximately
Fl2 to Fl5;, of an amphiidiploid derived fi-rom tihe cross Phascolus lunatu
L. x P, po~lystachvus~v~ l (L.) DSP. wJas studied. I nves -i ga ti ons prIev i -
ously had been condlucted through the F16 generation,
Using standard cytological techniques, root tips w;ere fixedl,
sectioned and stained with crystal violet-iodine stain to reveal the
A method was devised for concentrating beani buds inito small par--
affin blocks for microtome sectioning. Thle buds wer-e packed in Tissue-
mat just above the melting point in I cm. sections of plastic soda
straw~. The sections wer-e held erect. inside a plastic vial cap during
the infiltration process. After hardening ini an ice bot~h the str-aw
sections wrere separated, thie straw oncasements wer~e removed and thle
small cylindrical blocks wer-e mounted on a micr-otome block. Two oppo-
site sides of the cylinder were shaved parallel to the knife edge For
Standard methods wrere used for- anthorci squiashes with acetocar-mine
sta7ining. DiFficullties wocra encoun~tclerd with, this techniiqrue due to
the small size of the buds and the scarcity of pollen moth-er cells
exhibiting suitable division stages.
Pollen size studies were made using pollen whichh had been de-
posited on the hairs of the stylar brush at anthesis.
In the amphidiploids, melotic abnormalities included laggard
chromosomes, groups of chromosomes, chromatin strands and bridges.
In gener-al, st~erl'lity and fertility wer-e correlated with- the extent
of abnormal m~elotic figures.
Variation1 from the norm in pollen graiin size was takeni to be. the
result of chromosome elimination or duplicationi at meiosis. In at
least one case there seemed to be a correlation between the fertility
of the plant as measured by seed production and a narrow! range of
variability in pollen grain size. This is regarded as evidence for
diploidization in this particular plant.
Melotic disturbances were associated w~ith heterogenetic pairing
and cr-ossing over between interspecific homnoeologues. This could ex-
plain the deficiencies, duplications and chromatin bridges which w~ere
On the basis of these observations, continued selection should
increase fertility through the selection of balanced chromosome types
and the elimination of unbalanced chromosome combinations.
Although diploidization seemed to be progressing in the above
mentioned case, there was no evidence of a return to the 2n=22 state
of the diploid parents. All obser-vations of suiitable meiloic stages
indicated that the chr-omosome number was at or near t~he 2n=44 level
expected in the amphidiploids.
The Leguminosae is one of the mIoSt important families of flow~er-
ing plants from the standpoint of both economic and nutritional value
(41, 60). The lima bean, Phaseolu~s lunatus L., is a vegetable legume
which had an estimated value of 25 million dollars from 118,200 acres
harvested in the United States in 1968 (1). This represents an in-
crease in market value since 1954 of approximately 6 million dollars
per year- with a decrease of about 12,000 acres per year in the har-
vested acreage. The lima bean, like practically all edible legumes,
also represents a substantial source of vegetable protein and other
nutrients in areas of the world where it can be grown (25, 41, 60).
Injury to the apex, cotyledons, or hypocotyl when the seed germi-
nates in crusty or compacted soil is one of the problems of lima bean
production. The epigeal germination exhibited by present commercial
varieties results in this mechanical damage. The uniformity and stand
of the planting are consequently reduced.
Several studies (12, 13, 30, 44, 69) have been made using inter-
specific hybridization to incorporate a hypogeal germination character
into a species with epigeal germination. The hypogeal character would
eliminate many injuries due to cotyledon or hypocotyl breakage since
only the plumule and epicotyl would elongate and break the surface of
The cross P. lunatus var.'Fordhook'x P. polystachyus was first
made by Lorz (44). P. pol~ystachyus exhiibits the hypogeal germination
character and, being a wild native species, may exhibit other desirable
characters such as disease and insect resistance, vegetative vigor, and
better area adaptation.
Fozdar (19) reported on the cytogenetics of the parents, the
highly sterile FI, and the fertile amphidiploid derivatives from the
Dbaliwal et al. (16) reported on the abnormalities found in the
highly sterile FI of the P. lunatus x P. polvstachvus cross. Seve ral
melotic irregularities appeared to be responsible for high pollen
Stall (64) reported on the cytogenetics of a suspected triploid
derived from a backcross of an amphidiploid with a highly heterozygous
P. lunatus seed parent. He found this progeny to be a tetraploid and
speculated that it arose by fertilization of an unreduced (2n) female
gamete. The chromosome complement of the tetraploid would theoretically
be LLLP, letting L stand for a lunatus genome and P stand for a poly-
Lorz (47) has maintained the phenotypes of the amphidiploids
described by Fozdar (19) and Stall (64) and has made selections in
subsequent generations for increased fertility and productivity within
each phenotype. The lines are now estimated to be in the Fl2 to Fl5
generation. Selections have been made with the long-range goal of
developing a commercially acceptable variety with characteristics
superior to present diploid varieties.
Another aspect of the lima bean breeding program is the possibility
of backcrossing the amphidiploid to a P. lunatus type to obtain a
triploid individual. The triploid FI then might be further backcrossed
to P. lunatus to establish a series of aneuploids which would consist
of 2 P. lunatus genomes plus individual pairs of homologs of P. pol~y-
This study is concerned with the cytogenetics of amphidiploid lima
beans after several generations of breeding and selection. A study of
pollen grain sizes and their frequency distribution was made for a
comparison with similar studies made earlier by Fozdar (19) on earlier
amphidiploid generations. This might define more closely the reasons
for observed genetic variation and changes in fertility. Such informa-
tion would be instrumental in determining the feasibility of continuing
certain lima bean lines and discarding others from the breeding program.
The term am~phidiploid refers to plant forms which possess the
diploid chromosome complement of both parental species in their somatic
cells (23). Since the formulation of Winge's hypothesis in 1917,
amphidiploidy in higher plants has had far-reaching significance in
regard to theories of evolution and orgin of new species. This is
especially pertinent in view of the fact that at least half of the
species of known angiosperms of known chromosome number are polyploids
(22), According to Clausen and Goodspeed (ll), Winge's hypothesis
states that interspecificc hybridization followed by chromosome doubling
leads to tetraploids which are essentially homozygous diploids which
remain fertile and constant." This hypothesis has been substantiated
many times and is now a well accepted mode of origin of new plant
Most amphidiploids have been produced experimentally although a
few have been found to occur in nature. Those conditions once used to
induce experimental tetraploidy such as decapitation (34, 38, 42),
temperature shock (17, 53, 55, 58), and nutritional deficiencies (70)
are thought to be the cause of chromosome doubling in nature. Accord-
ing to Goodspeed and Bradley (23) the hybrid state itself may be the
most important factor in determining whether an amphidiploid arises
naturally from an interspecific cross. The absence of pairing chromo-
somes appears to induce a high percentage of restitution nuclei which,
in turn, leads to the doubled condition upon fusion of these restitu-
Huskins (31) used cytological evidence to confirm the theory that
Spartina townsendli had arisen from crossing S. alterniflora x S.
stricta. The chromosome number of the hybrid was 12.6, the sum of the
somatic chromosome numbers of its suspected parents. Other evidence
pointing to its origin in interspecific hybridization included high
fertility, true-breeding habit, and the formation of bivalents at
meiosis. It also demonstrated greater size and vigor.
Spartina townsendli should be considered as the classic example
of the evolutionary importance of allopolyploidy. The amphidiploid
is reported to have eliminated almost completely its parental species
from areas in which it was found, indicating an improved survival value
over that of its parent species. It can only be speculated as to how
many more common plants have had their origin as an interspecific
hybrid and, through natural selection, eliminated their parental
species from the ecological scene.
Clausen (10) observed that Penstemon neotericus represented a
recombination of most characters used to distinguish P. laetus and P.
azureus from each other. The range of PI. neoterious was bordered by
both of the supposed parental species. The hybrid apparently did not
backcross with either parental species.
Penstemon neotericus was found to have a chromosome number, n=32.
All forms of P. laetus had a chromosome number, n=8, while two forms
of P. azureus growing in the vicinity of P. neotericus had a chromosome
number, n=24. Polyploidy was rare in Penstemon, and the sum of the two
supposed parental species equaled the chromosome number of P.
neotericus. This would be entirely out of the ordinary unless it
originated as an interspecific hybrid.
Muntzing (50, 51) obtained an interspecific cross between
Galeopsi puesens (n=8) x G. speciosa (n=8). In the F2 a triploid
appeared, apparently due to non-reduction of a gamete. When the trip-
loid was subsequently backcrossed to Gi. pubescens, non-reduction again
occurred, this time in the triploid, and a 3n gamete was produced.
The offspring of the cross w~as found to be an amphidiploid with 32
chromosomes in the diploid state. The amphidiploid morphologically
resembled a species found in nature, G. tetrahit, which also had 32
somatic chromosomes. The artificial Gi. tetrahit crossed readily with
the natural G. tetrahit and meiosis was normal in the hybrid offspring.
This work exemplifies a method of modern evolutionary research in which
an existing polyploid species was resynthesized experimentally.
Clausen and Goodspeed (ll) crossed Nicotiana glutinosa (n=12) x
N. tabacum var, purpurea (n=24) to obtain a partially fertile hybrid,
N. dilt (n=36). Bivalents separated regularly in the FI while
univalents lagged in the equatorial zone during meiosis. -Pollen was
completely shriveled and devoid of contents. Tetraploids found in the
F2 were fertile and exhibited normal meiosis. The pollen grains were
approximately twice the volume of parental pollen grains. They proposed
that a tetraploid hybrid may be formed in a variety of ways: 1) by
doubling of the chromosome number immediately subsequent to fertiliza-
tion, 2) by bud variation in an FI interspecific hybrid, 3) by crossing
together tetraploid representatives of two different species, or 4) by
irregular distribution of chromosomes in an interspecific hybrid in
which the chromosomes do not pair in meiosis,
Karpechenko (37) presented a classic work in the field of crossing
distantly related plants. The intergeneric progeny, Raphanobrassica,
contained 18 chromosomes, 9 coming from each of the two parents,
Raphanus sativus L. and Brassica oleracea L. The FI was highly sterile
and only univalents were present at metaphase 1. A few seeds wlere
produced from the Fl. These fertile F2 plants proved to be tetraploid
with 36 somatic chromosomes. It was presumed that fertility was re-
stored by doubling of the chromosomes by fusion of unreduced gametes.
Bivalents were formed in meiosis of the F2.
Newton and Pellew (52) stated that Primula kewensis was found in a
group of seedlings of P. verticillata. This was later confirmed by
making the species cross and observing the offspring. Cytological
examination revealed that the parent species as well as the sterile
offspring possessed 18 somatic chromosomes. An occasional fertile form
was found which exhibited 36 somatic chromosomes. The chromosome sets
of _P. floribunda and P. ver-ticillata were found to be indistinguishable.
In the diploid hybrid and the tetraploid derivative, the same chromo-
some sizes and structures were observed. There were nine loosely paired
bivalents in the diploid hybrid at meiosis. Microsporogenesis was
apparently regular until late in pollen development when the pollen
degenerated and empty pollen grains resulted.
Newton and Pellew (52) stated that doubling of the somatic chromo-
somes resulted in a chimera of fertile tetraploid tissue. In the
tetraploid hybrid there wrere usually 16 bivalents and one quadrivalent
in mretaphase although as many as three quadrivalents were observed.
They indicated that Winge's hypotheses was adequate to explain the
high degree of fertility and constancy of the tetraploid. They also
stated that occasional pairing occurred between unlike chromosomes
Lindstrom (4'2) studied the inheritance of fruit size in tetraploids
of the cross .Lo~~LEResiu pimpinellifolium~! x L. esculentm. The F1
and the F2 of the cross were fertile even though the chromosomes of
the wild species were 30 percent smaller than those of the domestic
form. Tetraploidy was achieved by decapitation of F~ seedlings. Eight
percent of the vegetative shoots arising from the callus were found
to be tetraploid. The high fertility found in the tetraploid was
thought to be due to the strong tendency toward bivalent chromosome
association in meiosis.
Kostoff (39) crossed Nicotlana gqlauca Grah. (n=12) x N. langsdorffi
Weinm. (n=9). In somatic tissues of the hybrid the chromosome number
was n=21 although some cells were noted with aberrant numbers. Several
forms of abnormal mitosis were also observed. In meiosis in the diploid
hybrid, variable numbers of bivalents and univalents were found.
Bivalents had to be due to allosyndetic pairing. Univalents usually
led to the formation of restitution nuclei and numerous micronuclei.
Crossing-over was found to occur in the diploid hybrid if allosyndetic
pairing occurred. On crossing the FI hybrid with N. lanqsdorffi, the
progeny exhibited 30 somatic chromosomes, interpreted as being 2
langsdorffi genomes and one glauca genome. Upon crossing the triploid
(di-lan~gsdorffi--mo-m no -qLalauca) with N. glauca, two amphidiploids were
were obtained due to formation of monads in the triploid. Meiosis in
the amphidiploids showed formation of bivalents, trivalents, quadriva-
lents, and univalents. The latter combinations led to abnormal meiosis
and pollen abnormalities. In advanced generations, the progeny of the
amphidiploid exhibited increased pollen viability. By the Fg generation,
99.5 percent of the pollen was apparently viable. The number of multi-
valents and univalents decreased with increasing numbers of generations
while fertility increased. The amphidiploid was not stable and segre-
gated to give plants differing in leaf type, flower shape and size,
growth habit, and number of chromosomes,
Kostoff (39) stated that inconstant amphidiploids may give rise to
a series of adaptable forms. In certain cases they may afford more
suitable material for natural selection than highly constant amphidip-
Jones and Kobayashi (33) reported crossing Ipomoea pes-caprae ssp.
brasiliensis (L.) van 00ststr. x I. crassicaulis (Benth.) B. L. Robinson.
The six hybrids obtained were all morphologically different but were
intermediate to the parent species. Backcrosses to the parent species,
intercrosses between the hybrids, and selfs on the hybrids all resulted
in no seed set, Parent species set seed readily but no open-pollinated
seed were set on the hybrids. Chromosome counts showed 30 somatic
chromosomes in both parent species and the hybrids. Regular meiotic
behavior was found in both parents. The hybrids exhibited little pairing
of melotic chromosomes with zero to nine bivalents being formed, In
later stages lagging chromosomes and other abnormalities were frequent.
The tetrad stage was quite irregular with two large dyads being prevalent.
Pollen grains of variable sizes resulted. Pollen grains appeared viable
at anthesis but were not functional presumably due to an unbalanced
chromosomal complement. They suggested that stability of pollen size
be used as a measure of pollen fertility rather than the appearance of
stained pollen and concluded that the species were somewhat related be-
cause of some chromosome pairing at metaphase 1.
Fertility Changes in Auto- and Allotetraploids
in dealing with the problem of changes in the fertility of amphi-
diploids due to evolutionary changes in the genetic or chromosomal
complement of such amphidiploids, several cases can be cited of such
changes known to occur in autotetraploids. This comparison is particu-
larly valid for those amphidiploids considered to be segmental allopoly-
ploids which synapse at several points along the chromosome.
Sparrow et al. (63) found that differences in fertility of auto-
tetraploids derived from different varieties of snapdragon were associ-
ated with the frequency of lagging chromosomes and other abnormalities
seen at later stages of meiosis.
Bremer and Bremer-Reinders (8) found in autotetraploid rye that
different combinations of quadrivalents, trivalents, bivalents, and
univalent chromosomes were formed in the first metaphase. During ana-
phase, this gave an irregular distribution of chromosomes to the
daughter nuclei. The sex cells may have been fertile and, upon re-
combination, may have formed aneuploid individuals. Most aneuploid
rye individuals showed a low degree of fertility as well as poor vege-
tative vigor. By selecting tetraploids in each generation, the per-
centage of aneuploids produced in the following generation could be
reduced. Over a seven year period an increase in fertility was found
that may be correlated with an increase in regularity of meiosis.
Bremer and Bremer-Reinders (8) stated that when t'rivalents or
univalents were formed at metaphase 1, the remainder of the meiotic
process was likely to be abnormal. Usually two chromosomes of a tri-
valent reached one pole with one chromosome going to the other pole.
With univalents three things occurred: 1) a univalent reached a pole
and entered into the daughter nucleus in an undivided state, 2) the
univalent formed a micronucleus if it did not reach a pole, or 3) the
univalent may have lagged behind in the spindle and divided into two
chromatids. The chromatids reached the poles after the whole chromo-
When melotic division in tetraploid rye was regular there were
14 chromosomes in metaphase 11 that were split in anaphase II. In
the case of irregular melotic division, chromatids were frequently
present in metaphase 11 plates, derived from the splitting of univalent
chromosomes in anaphase I. These chromatids were incorporated into a
tetrad nuclei, or more frequently formed micronuclei. The frequency
of micronuclei appeared to be a good measure of the degree of meiotic
irregularity. Improvement in seed setting was found in later genera-
tions and was thought to be related to increases in melotic regularity.
Gillis and Randolph (22) stated that autopolyploids of recent
origin typically formed multivalents due to random pairing of chromo-
somes which were sufficiently homologous to synapse freely. Allopoly-
ploids ordinarily formed bivalents exclusively because the chromosomes
of the species concerned were sufficiently unlike to limit pairing to
the reduplicated parental genomes. The presence of quadrivalents in
a natural tetraploid could be interpreted wlith reasonable assurance as
evidence of an autoploid or essentially autoploid origin. Absence of
quadrivalents was not acceptable proof of the allopolyploid origin
for two reasons: 1) some autototraploids with small chromosomes formed
bivalents, and 2) multivalent association of chromosomes in autoploids
may not persist indefinitely but many disappear and be replaced by
regular bivalent synapsis.
The relative frequencies of quadrivalent and bivalent formation
over a ten year period in autoploid maize were studied by Gillis and
Randolph (22), They found fewer quadrivalents and more bivalents at
the dikinesis stage at the end of ten years than at the beginning.
This suggested that autoploids which formed multivalents with high
frequency at their origin may shift to the bivalent type of synapsis
by evolutionary means.
Roseweir and Rees (56) also believed that the fertility of auto-
tetraploids depended on the types and distribution of chromosome asso-
ci.ation at meiosis and proposed that selection for heritable change in
the distribution of the associations should be effective in improving
fertility. They found that fertility of autotetraploid rye was depen-
dent on the pattern of chromosome association at meiosis and that such
association was under genetic control. Selection for high fertility
in rye was based on increasing the quadrivalent frequency and reducing
the trivalent frequency.
Hilpert (27),studying tetraploid rye, found the chromosomes to
behave as quadrivalents, trivalents, bivalents and univalents. The
trivalents and univalents as well as the ring quadrivalents caused
melotic irregularities and the occurrence of aneuploids, Strong selec-
tion for high seed set and tillering for three generations resulted
in a significant increase in meiotic regularity.
Durrant (18) stated that in tetraploids the degree of mutivalent
formation at meiosis was often used as a criterion of homology between
the chromosome sets and gave information on the genetic control of
chromosome association. Formulas and tables of association frequencies
in tetraploids were presented.
Sybenga (68) stated that autopolyploidy resulted in reduced
fert ili ty. Sterility was due to a physiological inbalance because of
the duplicated nature of polyploids or to the high risk of unequal
separation of multivalents. He further stated that homoeologouis chro-
mosomes of amphidiploids may have been identical at sometime but have
become differentiated through evolution. Homoeologous chromosomes would
not have the opportunity to replace each other. Consequently, heterosis
was maintained if the homoeologous pairs carried heterozygous genes.
The author suggested that autotetraploids could be presumed to give
rise to allotetraploids by means of chromosomal rearrangements which
resulted in the inability of homoeologous chromosomes to pair at all.
In this manner, heterosis could be maintained if genes were heterozygous
at both loci. Fertility could also be improved since the chromosome
complement would behave as an allopolyploid.
Grum (24) studied melotic pairing in interspecific hybrids of Poa
under different environmental conditions. He found an unexplained in-
crease in univalent frequency in certain culms of a particular plant.
An increase in univalent frequency was also noted in plants grown under
alpine conditions as compared with plants grown at lower elevations.
He concluded that plant sensitivity to the environment resulting in a
physiological weakening of the plant was responsible for the increased
frequency of univalents at meiosis.
Yarnell (71) studied tetraploids which arose from the cross
Fra aria bracteata Heller x F. vescer rosea Fost. During meiosis an
occasional lagging pair was observed in anaphase i. At anaphase II
laggards were also observed but micronuclei were rarely formed. Most
F3 plants showed increased regularity over the F2 although some Fj
plants showed an increased amount of irregularity. Chromosome elimina-
tion was observed in the tetraploid. Quadrivalents were observed at
Shaver (61) found that preferential pairing and segregation in
allopolyploid maize was strongly influenced by the presence of an
inversion. It was theorized that in nature an amphidiploid containing
an inversion, a "structural hybrid," would be a positive force in
diploidization of the allotetraploid. This would be due primarily to
the abolition of interspecific crossing-over in the inverted region.
Chromosome bridges observed in the study did not appear to be typical
inversion bridges but appeared to be misdividing, bridge-forming
Beasley (6) observed that in species hybrids in Gossypium the
amount of pairing varied from complete to almost none. In Gossypium
hybrids with reduced pairing there was evidence that structural differ-
ences existed between the chromosomes. Most of the chromosomes formed
bivalents in polyploids that were produced from hybrids with a reduced
amount of chromosome pairing.
Stebbins (65) stated that a large proportion of the structural
differences between genomes of diploid hybrids involved small segments
that do not give typical meiotic configurations. Diploid hybrids which
are sterile but display regular meiosis may best be described as having
cryptic structural hybridity resulting in near sterility. Tetraploids
from such hybrids should exhibit complete pairing of the chromosomes.
The typical alloploid is usually derived from a wider interspecific
or intergeneric cross.
Stebbins (65) recognized two general types of allopolyploids;
typical, in which there was no intergenomal pairing and the identity
of the parental genomes was preserved, and segmental, in which inter-
genomal pairing was partial or complete, and the identity of the
parental genomes was lost. The author suggested that the terms homo-
genetic and heterogenetic be used in referring to the pairing of chro-
mosomes at meiosis. Homogenetic pairing should be used in describing
pairing in diploids or autoploids. In FI hybrids between two distinct
diploid species, only heterogenetic pairing can take place. A hybrid
between two distinct allopolyploids can have two different types of
heterogenetic association, namely, allosyndesis, or pairing between
chromosomes derived from different parents, or autosyndesis, pairing
between the different genomes derived from the same parental gamete.
Stebbins (65) stated that heterogenetic association is known to be
responsible for genotypic aberrations in otherwise true breeding allo-
polyploids. Even a small amount of this type of pairing usually leads
to some sterility as wlell as inconstancy. Its absence in established
allopolyploids is probably due to selection away from this type of
pairing. This is termed "diploidization," or the progressive change
of behavior to that of a diploid. In self-pollinating plants, diploid-
ization results in establishment of numerous cytogenetically different
Stebbins (65) explained chromosomal sterility in allopolyploids
in the following manner. The presence of bivalents formed by hetero-
genetic association in a diploid hybrid results in reduced as well as
unreduced gametes. The union of two unreduced gametes results in an
allopolyploid with considerable structural hybridity and subsequent
chromosomal sterility. This is a common occurrence in later generations
of amphidiploids produced by somatic doubling of a diploid hybrid.
However heterogenetic association may be eliminated through differential
affinity of the chromosomes and continued selection.
Segmental allopolyploids closely resembled one or both parents
because of segregation of interspecific differences. Cytologically
they demonstrate multivalents at meiosis in varying numbers. The
segmental allopolyploid would not be expected to breed true and segrega-
tion of morphological characteristics in early generations permits
selection and establishment of a series of different lines. This also
permits selection of adapted types with high survival value in early
Stebbins (65) stated that differential affinity is a characteristic
of segmental allopolyploids. Chromosomes pair by segments and the more
extensive the homology of the two chromosomes, the greater the affinity
for pairing. Sterility in diploid hybrids which have nearly complete
pairing at meiosis may be due to random segregation of small non-
homologous segments. This results in gametes with nonviable duplica-
tions or deficiencies. Sterility is partially overcome in allotetra-
ploids from such hybrids due to homogenetic pairing.
Interspecific Hybridization in Phaseolus
Allard and Allard (2) stated that eleven species of beans of the
genus Phaseolus are known to be cultivated in tropical, subtropical,
and warm-temperature regions.
Hedrick (26) reported that the lima bean is a native of South
America where it grows as a perennial. It was widely grown as a garden
vegetable in the eastern states after 1825. The origin of the lima
bean was more closely pinpointed by Mackie (48) and Seelig and Roberts
(59) as being in Guatemala.
Allard and Allard (2) also stated that P. polystachyus occurs
from Connecticut southward to the Gulf of Mexico. Although stated
otherwise in Gray's Manual of Botany, a characteristic of its germina-
tion is the hypogeal type of germination while most other species have
epigeal germination. The Scarlet Runner bean (P. multiflorus = P.
coccineus) also exhibits the hypogeal germination character.
The 2n chromosome number is 22 in all the species (14, 15, 16, 19,
P. polystachyus is unsuited to tropical day lengths of 12 hours or
less and to winter conditions of daylight in the greenhouse. Growth
is slow and the plant becomes dwarfed and produces few flowers under
these conditions. The rootstock is perennial, the stems dying down
during the winter. Buds are formed which produce the new stems the
On the basis of the constant chromosome number of the species of
Phaseolus, Allard and Allard (2) believed this genus to be a relatively
stable one. It might also be assumed that it is of more recent evolu-
Strand (66) crossed the Urd bean, P. munqo with P. vulgaris to
secure resistance to Mexican bean beetle in a commercially desirable
type. It was implied that a backcross to P. vulgaris using Fl pollen
was not successful because of environmental conditions. The cytogenetics
of the FI were not investigated. The Fl, F2, and F3 were partially
fertile and produced several seed.
Lamprecht (40) found that P. vullifis and P. multiflorus differed
by 2 main genes; epigeal or hypogeal germination, and the stigmatic
surface running either down the inner or outer face of the pistil.
Neither of these genes could be transferred to the other species because
a gametic or zygotic lethal was produced.
Gates (21) reported that P. vulgari has epigeal germinationl while
P. coccineus; exhiibits h-ypogeal germination, and that these. characters
are controlled 'Oy a pair of interspecific 9enos. He postulated that
the ancestors of th-ese two species were heterozygous for this pair of
genes and the species originated through the mechanisms of' self--sterility
and mutation. P. coccineus still exhibits the self-sterility character.
Lorz (44) was the first to report the cross P. lunatus L. x P.
polystachyus (L.) B.S.P. The cross was successful when P. lunatus was
used as the seed parent. All hybrids expressed the hypogeal germination
habit of P. polystachyus. The principal objective of the cross wias to
incorporate the hypogeal character into a commercial P. lunatus type
to solve the problem of mechanical injury to the hypocotyl and cotyledons
during emergence in compact or crusted soil. He also su99ested that
P. polystachyus, being a wild native species, may possess resistance
to certain diseases and insect pests and have physiological properties
which would enhance its natural survival value.
Honma (28) made the interspecific cross P. vulaSELs L. x P.
acutifolius Gray to incorporate common bean blight resistance into
P. vulsaris. The hybrid wias shorter than either parent. Foliage of
youn9 plants resembled P. acutifolius but,with a e, looked more like
P. vulgaris. The hybrids were fertile and set seed. Cytological in-
vestigation of root tip preparations did not suggest marked morphological
differences between either of the two species and the interspecific
Baggett (5) made crosses between P. vulgaris and P. coccineus in
an effort to obtain resistance to bean yellow mosaic virus since P.
coccineus was completely immune to a large number of BYMV strains. The
FI was vigorous, uniform and intermediate to the parent species in
many plant characters. Seed production was lower than either parent
indicating possible genetic or cytological incompatibilities. The F2
included a great diversity of plant types with all degrees of fertility.
Buishand (9) used P. dumosus, a species native to Central America,
as the pollen parent in crosses with P. vulgaris. Hedrick (26) reported
that P. dumsu is probably a form of P. lunatus, The cytogenetics of
the hybrids were not reported.
Wall and York (69) crossed P. vulgaris with P. coccineus to deter-
mine the inheritance of seedling cotyledon position. They reported that
P. vulgaris had epigeal and P. coccineus had hypogeal germination. P.
coccineus could not be used as the female parent in the cross. A distri-
bution curve indicated that position of cotyledons was conditioned by
multiple factors. Fertility of the FI was sufficient to produce an
adequate F2 generation.
Lorz (45) suggested that the oriental species of Phaseolus: aureus,
munqo, angularis, aconitifolius, calcaratus, and radiatus have floral
morphology, climatic adaptation, and plant habit with a stronger affinity
to the genus Vigna than with P. vulgaris or P. lunatus. Hle stated that
the possibility of transferring characters from these species to occi-
dental species by interspecific hybridization appears unlikely.
Lorz (46) also reported the attempted interspecific crosses in
the genus Phaseolus and the success attained with each. Listed as
successful were P. vulgaris x P. coccineus, P. vulgaris x P. polyanthus,
Pl. vulgaris x xanthotrichus, P. vulgaris x P. qclabellus, P. vulgaris x
P. acutifolius, _P. lunatus x P. polystachyus, and P. lunatus x Phaseolus
sp. (P,I, No. 201103). Those listed as unsuccessful were _P. vulgiaris x
P. lathyroides, P. vulgaris x _P. atropurpureus, P. y Offris x P.
aureus, P. vulgaris x P. munqgo, P. vulgaris x P. angqularia, P. vulgaris
x P. calcaratus, _P. vulgaris x P. helvolus, P. vulgaris x P. speciosus,
and P. vulgaris~ x _P. lunatus.
Lorz (46) stated that the cross P. lunatus x P. polvstgging was
successful when P. lunatus was used as the seed parent, The main
objective was the introduction of hypogeal germination into the P.
lunatus type to prevent "bald heads" and breaking of the hypocotyl
Honma and Heeckt (30) made the cross P. lunatus L. x P. coccineus
in an effort to transmit the hypogeal germination character into the
lima bean. The FI exhibited the hypogeal germination character.
Vegetative characters of the FI were more like P. coccieus than P.
lunatus. Several seed were produced on the FI plants indicating mod-
erate fertility. In the F2 the germination habit ranged from epigeal
to hypogeal. Other characters of the F2 resembled parental types or
recombinations of parental types.
Honma (29) crossed P. vulqaris and P. lunatus in order to incor-
porate the green cotyledon character into P. vulgaris. Most of the Fl
plants resembled _P. vulgaris in morphology. Pod and seed shape were
intermediate between the two parents, Fertility in the F2 and F~ varied
from normal to no set. Sterile types ? re generally abnormal in vege-
tative characters. Some of the hybrid could be backcrossed to both
parents while others failed to set seed". The use of gametic diversity
aided in making the interspecific cross.
Kammermann and Bemis (35) crossed the Scarlet Runner bean, P.
coccineus L., with the snap bean, P. vu~lqaris L., in an effort to
explain the seedling irregularities inherently present in the FI plants.
Two types of dwarf plants were reported to occur in the FI, B-dwlarfs
which died in the seedling stage when the variety 'Blue Lake' wras used
as the P. vulgaris parent, and T-dwarfs which grew to maturity when
the variety 'Ten~derareen' was used as the P. vulgaris parent. Appar-
ently viable pollen was produced by T-dwarfs but no seeds were obtained.
When an intraspecific P. vulqaris F, ('Blue Lake' x 'Tendergreen') was
used as a parent in the cross with P. coccineus, four types of offspring
were produced in approximately equal numbers: B-dwarfs, T-dwarfs,
BT-dwarfs, and normals. The normal FI produced viable seeds when polli-
nated with P. vulgaris, P. coccineus, or sib pollen.
Bemis and Kedar (7) crossed five varieties of P. vulgaris with a
common variety of' P. coccieus to study the inheritance of abnormal
seedling development. They reported the cross easy to make but that the
result wras numerous seedlling abnormalities. They proposed a series of
three alleles at twro independent loci to explain the inheritance of
Fozdar (19) made cytological investigations of FI, backcross and
amphidiploid derivatives of' the cross P. lunatus L. x P. polystachyus
(L.) B.S.P. Some morphological characters of the parental chromosomes
were similar while others were distinctly different. It was noted that
PI. polystacyu chromosomes were larger than those of P. lunatus and
that the difference in pollen grain size may have been due to the dif-
ference in chromosome size. The FI hybrid showed several melotic
abnormalities such as presence o-f univalents, laggards, chromosome
bridges and fragments. The amphidiploid F2 had a somatic chromosome
number of 44 as did the Fj and Fq progeny with the exception of one
aneuploid plant which apparently had 42 somatic chromosomes. Quad-
rivalents were formed in the amphidiploids and led to some melotic
Dhaliwal et al. (16) reported that the hypogeal character was
intermediate in the F1 of the interspecific cross P. lunauis. L. var.
'Fordhook' x _P. ~po4l.s~tachyus (L..) BS.P.
Cytological Investigation indicated abundant irregularities in
most pollen mother cells of the Fl. Several univalents were observed
at metaphase I and may have been due to failure of pairing at prophase I.
Q~uadrivalents were also observed. Chromosomes at metaphase I formed a
mass in which individual bivalents could not be distinguished. Such
an irregularity has been attributed to disturbed nucleic acid metabolism
(13). Laggards and anaphasic bridges were observed at anaphase 1.
Pollen was variable in size and most was sterile, Meiotic irregular-
ities seemed to be responsible for high pollen sterility.
Al-Yasiri and Coyne (3) obtained the interspecific cross P.
vuiar~is xP. acutifolius aided by the use of growth regulators naph-
thalene acetamide and potassium gibberellate applied to the pedicel of
pollinated flowers. The fertility or cytogenetics of the hybrid were
Coyne (12) attempted to create a gene pool in the genus Phaseolus
by crossing P. acutifolius x P. coccineus. The cross wJas successful
only when P. acutifollus wras used as the female parent. CreatIi on of
the gene pool of these species, plus P. vulgaris, would enable plant
breeders to combine tolerance to drought and a wide range of bacterial,
fungal, and viral diseases in beans. Cotyledon position in the hybrid
was intermediate on the stem in comparison to the parents'. The flowers
of the hybrid were not self-fertile, and no success was obtained in
backcrossing the hybrid to either parent. Pollen production was
reduced in the hybrid but a high percentage of'the pollen appeared
Al-YEIsiri and Coyne (4) stated that the success of interspecific
hybridization in Phasolu species depended on parental heterozygosity,
the parent varietal lines used, and environmental conditions at the
time of pollination. They crossed various species of Phaseolus, all
having the same chromosome number (2n=22). Th ree cl as si f ica t ions wJere
established on the basis of seed and pod development. Compatible
crosses yielded hybrid seed. A compatible cross was P. vulqaris x _P.
cocci neus, Partially compatible crosses wlere where pods collapsed
during early stages of development. This class included P. vulgaris x
P. acutifolius, P. acutifolius x P. coccineus, P. _cocineus x P. acuti-
folius, _P. vulqaris x P. munqo, P. munqo x P. aLcalcaats P. vulqsris x
P. lunatus, P. anqularis x P. acutifolius, P. anqularis x _P. vulgaris,
-. munc0 x P. vulgaris, P. anqularis x P. calcaratus, P. vulSalis x P.
calcaratus and P. anqiularis x P. mungqo. Incompatible crosses wlere those
in which no pod development occurred. This class included all other
combinations of species crosses.
Stall (64) studied a plant derived from a cross between P. Tunatus
and an amphidiploid from a PI. lunatus x P. polystachyus cross, with
the view that the plant might be a triploid. Root tip smears indicated
that the plant was a tetraploid with 44 somatic chromosomes. Quad-
rivalents, trivalents, bivalents, and univalents were formed during
metaphase 1 of meiosis. Laggards, assumed to be P. polystachyus chro-
somes, were frequently observed.
MrATERIALS AND METHODS
Root Ti~p Sectioin
Seed of the lima bean (Phaseolus lunatus L. var. 'Fordhook') and
several F12 to Fl5 amphidiploid lima bean lines derived from the cross
P. lunatuis L. x P. polystachyus (L.) B.S.P. were germinated on moist
filter paper in petri dishes. Root tips I cm. long were clipped when
the radicle was 3 to 5 cm. long and again after secondary roots had
developed. Additional root tips were obtained after transplanting the
seedlings to soil in peat pots. A short length of nylon knitting yarn
was inserted through the bottom of the pot to serve as a wick and the
peat pot was suspended over a supply of water. The root tips were
collected as they grew through the peat pot. To facilitate handling a
large number of root tips during dehydration and paraffin sectioning,
the root tips were tied into bundles of 4 to 10 roots each as described
by Lorz (43).
A pretreatment of 2 hours in .002 M 8-hydroxyquinoline in the
refrigerator (20, 54) was used to cause metaphase arrest and to clarify
chromosome morphology. Fixation was in Craf fixative (57).
After 24 hours in the fixative the root tips were dehydrated and
embedded in paraffin for microtome sectioning according to some modi-
fications of the schedule given by Jenson (32). The tissue was left
in each member of the tertiary butyl alcohol series for 24 hours rather
than 2 to 4 hours. Final paraffin embedding was in Tissuemat with a
melting point of 52.20C
Root tip sections were cut on a rotary microtome at a thickness
of 15 microns. Ribbons were affixed to microscope slides with Haupt's
adhesive with 2 grams phenol added as a preservative (32).
The sections were stained with crystal violet-iodine according to
the schedule given by Sass (57). Before observation the mounting
medium was hardened for 2 to 4 days on a slide warmer at 40 C.
Flower Bud Sectioning
After a sufficient number of root tips had been collected, the
seedlings were transplanted into the field. A\ black plastic mulch was
used to control weeds. Wire and string trellises were provided for
the vining types and supplemental water was applied as needed by
Bud material for pollen mother cell observations was collected
from plants grown during the winter in the greenhouse and from plants
grown from seed and transplanted into the field. Collections were made
periodically between 7:30 A.M. and 10:00 A.M. It was found that certain
stages of meiosis could be found at any time during this period, but
that buds of the same apparent size would not always contain the same
Meiotic stages in pollen mother cells of lima beans are difficult
to find. The buds and anthers are small and few pollen mother cells
are found in each anther. In order to expedite the location of the
proper melotic stages for cytogenetic observation, the length of fixed
flower buds was measured with a vernier caliper after removal of the
subtending bracts and pedicle. The size range in which melotic stages
could be found was 2.7 mm. to 3.1 mm. Therefore, it was considered
necessary to continue to di sect buds within this size range since a
specific bud size for meiotic stages could not be defined with great
Flower bud material was pretreated, fixed and dehydrated in the
same manner as the root tip material. For more rapid penetration of
the pretreatment and fixative solutions, the buds were placed in a
vacuum oven and the pressure reduced to approximately .2 atmospheres
for 3 to 5 minutes.
Since lima bean buds are so small at the time wJhen meiosis is
occurring, embedding and sectioning one bud at a time was time consuming,
inefficient, and not always fruitful. Therefore a method was developed
in which several small buds in a paraffin block were sectioned at one
t ime, it was found that arranging buds in melted paraffin with a hot
needle was a useful procedure, but the resulting block was usually
quite irregular in shape and not suited for most efficient sectioning.
Only a few buds could be arranged in a clump for successful cutting.
Therefore, it was desirable to find a method of embedding in a small
container that could be peeled away, leaving a large number of buds
concentrated in a small paraffin block.
Melted paraffin was poured into a petri dish to a depth of approx-
imately I cm, The petri dish was placed on a hot plate and the temper-
ature adjusted to 560c. to 580C. by using a thermometer suspended in
the paraffin to measure the temperature. A plastic vial cap, 1.5 cm.
deep, wlas filled wlith melted Tissuemat and placed in the petri dish.
One centimeter sections of plastic soda straw were placed upright inside
the vial cap (Figure 1). The buds wlere dropped into the straw sections
with twleezers and tamped down with a stirring rod. No attempt was made
Figure 1.--Petri dish of melted paraffin containing plastic
vial cap wi th seven straw sections and buds to
be embedded. The thermometer was used to regu-
late the temperature of the hotplate.
to orient the buds in any particular direction within the straw section.
As many as seven straw sections could be placed in one vial cap and
filled with buds in one operation (Figure 2).
When the straw sections were filled to the desired depth, thle
plastic vial cap was carefully immersed in an ice water bath and allowed
to harden. The vial cap was flexible enough for the entire paraffin
block to be removed intact (Figure 3). The paraffin surrounding the
straw sections was cut away and a longitudinal slit was made in the
straw with a razor blade. The plastic straw peeled easily away from
the paraffinl cylinder containing the buds (Figure 4).
For sectioning, the paraffin cylinder was attached to a mound of
paraffin on a wooden microtome block by sliding a hot spatula between
the two pieces of paraffin and pressing the two melted surfaces together
(Figure 5). Two opposite sides of the paraffin cylinder which contained
the least amount of bud material were selected and cut parallel to each
other with a razor blade (Figure 6). Some bud tissues were normally
trimmed away in this operation but these were usually tissues external
to the anthers. No further trimming was necessary, and the block could
then be mounted on the microtome and a ribbon cut.
It was found that lima bean buds fixed and dehydrated in the manner
described tended to have hard areas in the buds that caused ribbon
sections to tear and split. Two procedures were attempted in order to
overcome this difficulty. After fixation in Craf fluid, the buds were
held for 15 minutes in IN HCI at 600C. for softening of the tissues.
This procedure improved the microtoming properties somewhat. When HCI
hydrolysis was combined with the use of Carnoy's II fixative (32), a
greater improvement was noted. Ribbons were generally much easier to
Figure 2.--Buds packed into soda straw sections arranged
in the plastic vial cap,
Figure 3.--Paraffin block before removal of excess paraffin
around the straw sections, and an empty plastic
Figure 4.--Paraffin cylinders formed in soda straw sections.
Left, straw intact; center, straw slit length-
wise; right, straw peeled away from the paraffin
Figure 5.--Paraffin cylinder attached to a microtome
block. Opposite sides of the cylinder have
been trimmed parallel to each other.
Figure 6.--Face view of trimmed paraffin cylinder showing
trimmed parallel sides and several embedded
cut without splitting, and longer ribbons could be cut using the same
microtome knife section.
Pollen Grain Study
Five newly opened flowers were collected from each amphidiploid
plant in the greenhouse and field plantings for a study of pollen grain
size distribution. A plant thought to contain three P. lunatus genomes
and one P. polystachyus genome (64), and two of its F2 progeny were
also included in the pollen grain study. The variety 'Fordhook' was
used as the diploid control.
The flowers were tripped and pollen collected from the hairs of
the stigmatic brush. The pollen from each flower was placed in a drop
of acetocarmine stain on a microscope slide and a cover slip applied.
No pressure was applied to the cover slip and the edges of the cover
slip were not sealed. Random pollen grains were chosen for measurement
by moving the microscope stage to the lower right hand corner of the
cover slip and scanning consecutive vertical fields across the cover
slip until 20 apparently viable grains had been observed. Measurements
were made by drawing two opposite and random dimensions of each grain
using an Abbe camera lucida. The average diameter of the two dimensions
was used for size determination of the grains. Twenty pollen grains
from each flower were measured, making a total of 100 measurements per
The 40x dry objective with 10x eyepiece was used for greatest
accuracy. Absolute size of the pollen grains was determined by measuring
the distance between twro points made by using a stage micrometer of
Although paraffin sectioning and crystal violet-iodine staining
technique gave good results for observing lima bean root tip chromo-
somes, the procedure could not be adapted for chromosome observations
in pollen mother cells of lima bean. Few cells were found which were
oriented properly for viewing meiotic figures at high magnifications.
The following alternative procedure, however, proved satisfactory.
Fixed flower buds were dissected in 70 percent ethyl alcohol and the
anthers squashed in acetocarmine stain. Anther tissues were removed
and a cover slip applied. The slide was heated gently over an alcohol
flame. Gentle thumb pressure with the slide between sheets of bibulous
paper was generally necessary to flatten the cells sufficiently to get
all parts of melotic figures into a single focal plane. Cover slips
were sealed with sealing wax and aged in the refrigerator for further
clarification of the chromosomes. The slides were not made permanent.
Photomicrographs, drawings, and other observations were made as soon
as possible after aging.
All observations were made using a Leitz phase contrast microscope
with 10x, 20x, and 40x dry and 90x oil immersion objectives. Eye
pieces were 8x and 10x.
Photomicrographs were made using the MIKAS 1/3x micro attachment
with a 35 mm. Leica camera. Kodak Plus-X Pan film was used for all
black and white photomicrographs and Kodak Kodachrome II color film
was used for color photographs.
RESULTS AND DISCUSSION
The plants grown in the greenhouse during the winter and those
grown from seed and transplanted into the field during the spring
provided a representative sample of the various phenotypes which have
been selected by Lorz (47) and described by Stall (64). Table I gives
a summary of characteristics of the diploid and amphidiploid plants
used in this study. Figures 7, 8, and 9 show the diploid variety
'Fordhook' and a bush and vining amphidiploid, respectively.
An unreported phenotypic variation was found in the amphidiploid
field planting. The two bush plants, numbers 6 and II, had very light
violet colored flowers as compared to the darker, more purple flowers
of several other plants. Lorz (47) stated that when this phenotype
was grown in the greenhouse during the winter, the pale violet color
faded out and was difficult to distinguish from the white phenotype.
The white, purple, and violet phenotypes are compared in Figure 10.
All amphidiploid seedlings, when transplanted to peat pots, still
displayed the hypogeal germination character. If the seed had been
planted in soil at a normal depth, the character might have been more
pronounced. Cotyledons at the surface of the soil can be seen in
For cytogenetic observation of small chromosomes such as those
found in root tips of lima bean, the use of paraffin sectioning appeared
to be a most efficient technique. As many as 72 paraffin sections,
each containing 10 root tip cross sections, could be placed on each
Flowering Root Flower Pod Seed Coat Seed Pod Plant
Plant Response Stock Color Size Color Productivity Habit
Characteristics of the diploid and amphidiploid lima bean
plants grown in the greenhouse and in the field, 1970
Flowering Root Flower Pod Seed Coat Seed Pod Plant
Plant Response Stock Color Si ze Color Productivi ty Habit
660; Day Neutral Pe renn ial Wh ite Med ium Wh ite Med ium Bush
985" Short Day Annual White Small Black High Vining
986` Short Day Perennial Whi te Small Black High Vini ng
"Grown in the greenhouse, early spring, 1970.
Table 1 (Continued)
. *' rrr -L;
Figure 7.--Diploid lima bean P. lunatus L. var.
'Fordhook', spring, 1970.
Figure 8.--Bush character of amphidiploid plant 6,
Figure 9.--Vining character of amphidiploid plant 9,
Figure 10.--Flower color of amphidiploid plants grown
in the field, spring, 1970.
Figure II.--Hypogeal character of amphidiploid seed-
lings transplanted into peat pots.
microscope slide. This gave a large amount of material concentrated on
a single slide. In addition, if care was taken in slide preparation,
all sections could be placed on the slide in serial fashion.
The squashi technique for root tip chromosome observation may give
immediate results more quickly than the paraffin sectioning technique,
but much more. time was usually spent with the squash method searching
for a cell wlith chromosomes in the proper orientation for counting,
drawing, or photomlcicrograph-y. With paraffin sectioning a polar view
of a metap~hase plate could normally be located rather quicklyl since
the morphology of the root cross section was not distiurbcd. It could
be ascertained at a glance if the tissue in question was likely to
contain meristemat~ic cells, whereas in a squash preparation the tissue
integrity w~as destroyed and one squashed cell looked, for- all pr-actical
purposes, like any other.
Table 2 presents the flower bud measurements of diploid and a:nphi-
diploid plants grown in the field and greenhouse and the stage found
at each bud size. These measurements represent buds collected at
various times during the morning hours and no collection time wJas
observed to have a greater meiotic activity.
The average bud size at which melotic figures were found was 2.9 mm.
However the size range within which meiosis occurred, 2.7 mm. to 3.1 mm.,
precluded discarding buds that fell within this range for fear of dis-
carding material containing m~elotic figures.
The bud size at which meiosis was found to occur in amphidiploid
plants did not agree with the findings of Fozdar (19) or Stall (64)
who stated that microsporogenesis takes place when buds are less than
2 mm. long. This discrepancy cannot be explained, except that perhaps
only estimates of bud length were made on their bud material.
Plant Length in mm. Stage~
Fordhook (2n) 2.9 Pollen
2 (4n) 3.1 M ,A ,T ,M l i
3 (4n) 3.0 Tetrads
2.9 PI, MI, AI
2,8 Early PM4C
4 (4n) 3.1 Pollen & Tetrads
3.0 Pollen, MI, TI,
2.8 Early PMC
5 (4n) 3.1 Pollen & Tetrads
2.9 Pollen & Tetrads
2.7 AI, TI, MIl, TII
6 (4n) 2.8 Pollen
2.8 Pollen, Tetrads,
MI, TI, MII, TII
2.7 Early PMC
9 (4n) 2.9 Pollen
2.8 Early PI
2.7 TI, Mil, TII
10 (rtn) 3.1 Pollen
2.9 Pollen & Tetrads
2.8 TI, TIl
Bud measurements of diploid 'Fordhook' and amphidiploid
lima bean lines and the melotic stage found in
anthers of each bud
Plant Length in mm. Stage
Table 2 (Continued)
Tetrads, TI, TII
Pollen & Tetrads
Pollen & Tetrads
Tetrads, TI, Mil,
M1, AI, TI, MII,
Pollen & Tetrads
PI, TI, MII, TII
MI, TI, MIl, TII
"PI-Prophase 1, Mi-Metaphase 1, Al-Anaphase 1,
MII-Mletaphase II, TIl-Telophase II.
With the wide range of bud sizes at which meiosis may occur in
amphidiploid plants, the selection of a single bud with the intention
of finding melotic figures was largely a matter of chance. Using the
soda straw embedding procedure, buds of the estimated proper size were
selected and concentrated within a small, compact and easy to handle
cylinder. A minimum amount of trimming was required, and the cylinder
could be mounted directly on the microtome block after removal of the
soda straw. Depending on the length and diameter of the straw section,
as many as 18 to 30 buds could be embedded in one cylinder.
Fewer microscope slides were required per bud using the soda straw
embedding procedure. Less time was required in the staining schedule
and more bud material could be concentrated per slide. This facilitated
scanning the anther locules for pollen mother cells and meiotic figures.
The use of paraffin sectioning for observing lima bean chromosomes
was not completely satisfactory. The anther locules containing pollen
mother cells undergoing meiosis could be found with ease under low
magnification. However, when an oil immersion objective was used for
greater magnification, the meiotic figures usually could not be focused
in a single viewing plane. The sections could hardly be cut less than
25 to 35 microns as recommended by Jensen (32) for fear of cutting
through an excessive number of division figures. Orientation of the
pollen mother cells to the cutting edge, then, becomes the main obstacle
to be overcome for effective use of the paraffin sectioning procedure
for lima bean buds.
The use of the squash technique for lima bean pollen mother cells
was found to be satisfactory. The cells could generally be flattened
enough to give a sufficient side or polar view of meiosis. Most anther
material had to be removed from under the cover slip for proper squashing
of thle pollen mother cells. The method described by Fozdar (19) was
found to be helpful in removal of excess anther debris.
Acetocarmnine was found to be an adequate stain for lima bean
chromosomes. Dbaliwal et al. (16) indicated that acetocarmine would
not stain lima bean chromosomes. Their results may have been influenced
by their use of propiono-alcohol fixative rather than Craf or Carnoy's
11 fixative. Clarity of the chromosomes was enhanced by aging the
slides in the refrigerator for 3 to 7 days after staining. The low
temperatures also permitted storage of the slides for several weeks
without the need for making them permanent.
The somatic chromosome number of the amphidiploid lima beans used
in this study was 44. This was confirmed in several counts of mItotic
chromosomes in root tip sections as shown in the photomicrograph and
camera lucida drawing in Figure 12. The somatic chromosome number, 44,
confirms the findings of Fozdar (19) and Stall (64) concerning the
~chromosome number of this amphidiploid material. Several of--these
chromosomes show similar morphological forms to those presented by
Fozdar (19). However, no constrictions could be seen, and all chromo-
somes could not be matched with the idiograms of Fozdar. This may be
due to different methods of pretreatment, fixation, and staining, and
the use of tertiary butyl alcohol dehydration and paraffin embedding.
The possibility also exists that the chromosome morphology of the
advanced generation amphidiploid plants may be altered through cross-
overs or translocations, particularly if allosyndetic pairing is
prevalent, as in a segmental allopolyploid.
Figure 13, a photomicrograph and a camera lucida drawing of the
Figure 12.--Above, phGOtmicrograph of a metaphase plate of a root
tip cell of amphidiploid plant 6 (1050%); below.), a
interpretive camera lucida drawing (2000X).
figure 13.--Above, photomriicrograph of a metfaphase plate of a3 root
tip cell of P. _tnz L. var. 'Fordhook' (1050%);
below, an interpretive camera lucid drawing (2000X).
same cell, shows a metaphase plate of a root tip cell of P. lunatus var.
'Fordhook'. Twenty-two chromosomes are present confirming the findings
of Karpechenko (36), Darlington and Ammal (14), Darlington and Wylie
(15), Dhaliwal et al. (16), and Fozdar (19). As with the amphidiploid
material, all chromosomes cannot be matched with the idiograms presented
by Fozdar (19) for P. lunatus, but several chromosomes appeared similar
Examination of several amphidiploid pollen mother cells obtained
by squashing anthers in acetocarmine revealed a meiotic chromosome
number of 22. This count is consistent with the results obtained by
Fozdar (19) and Stall (64) who made similar counts using earlier genera-
tions of the same material. Photomicrographs with corresponding camera
lucida drawings of amphidiploid pollen mother cells in three division
stages are shown in Figures 14 to 17.
Figure 14 shows 22 bivalents paired at diplotene, the melotic
stage indicated by the retention of a single nucleolus. Chromatids
are not apparent in this cell but synapsis has apparently lapsed, allow-
ing separation of the paired homologues. Chiasmata number and position
are variable between chromosomes.
Figure 15 shows a pollen mother cell with 44 chromosomes. In the
photomicrograph they may appear as a single plate but with careful
focusing of the microscope it could be seen that this was probably a
stage transitional between anaphase I and metaphase 11 since one plate
lies below the other. Each plate contained 22 chromosomes.
Figures 16 and 17 represent metaphase 11 of normal meiosis in
amphidiploid pollen mother cells. These cells are from plants 2 and
660, respectively. The camera lucid drawings show 22 chromosomes in
f ;1- .I;.
Figure 14.--Above, photomicrograph of a po~llen mother cell of
amphidiploid plant 660 at diplotene (1050X); 'oelow,
an interpretive camera lucid drawing (2000X).
Figre15,-Aov, potmicogap ofa olln oter el o
amphidiploid pln 6 tmtphs I(00)
below,~~~- an .-Intrrtv ae lcd rwn hwn
a( total of :A crmsmsi h twomtpaeI
Figre16.-Aove potoicogaph f pole mohe c ll o
amphidiploid~ pln tmtpaeI 15% ;blw
an~ ~ ~ ~ ~ ~~~. ineprtv camer ucddrwn(20X.\
Ir: i.- '
fi ~ r~r~i
i ,h s
Figure 17.--Aboie, photomicrograph
amphidiploid plant 660
below~, an interpretive
of a pollen mother cell of
at metaphase 11 (1050X);
camera lucida dr-awing
each plate of each cell. Chromatids can be distinguished in several of
As in the case of Dbaliwal et _al. (16), it was noted that individual
chromosomes usually could not be distinguished at late mietaphase i.
The chromosomes tended to clump into 3 to 5 groups and the integrity
of individual pairs was lost. Many camera lucida drawings reflect the
lack of resolution encountered at this stage. Therefore, only the
general outlines of the chromosome groups could be presented.
Most melotic figures, like those presented thus far, appeared
normal in all plants which were studied, both those grown in the green-
house and those grown in the field. Pairing and disjunction were normal
with bivalents being formed at metaphase I as in a natural diploid or
a typical alloploid. Bivalent pairing could be attributed to differen-
tial affinity in which parental chromosomes would tend to pair with
each other in preference to a homoeologous chromosome from the other
species. However, evidence supports the conclusion of Fozdar (13) that
progeny derived from doubling of the original diploid hybrid are
segmental allotetraploids. The evidence for this conclusion may be
summarized as follows:
In the FI the most frequent melotic pairing configuration at
dikinesis and metaphase I was 8 bivalents with 6 univalents. Many of
these bivalents may be only loosely held, as evidenced by the presence
of several pollen mother cells having fewer bivalents and more uni-
valents. Some segments of interspecific homoeologues must have had
closely related genes to cause segmental pairing.
In the F2, F and Fq amphidiploid progeny all pollen mother cells
did not show regular homogenetic pairing giving rise to 22 bivalents.
About 50 percent of the cells had quadrivalents. Trivalents and uni-
valents were also present. Quadrivalents may be attributed to secondary
associations between bivalents, again reinforcing the possibility of
identical segmenits on hom~oeologous chromosomes.
The karyotype of both P. lunatus and P. polystachyus is similar
enough to suspect some interspecific homology.
Genetic evidence also supports the thesis that the plants are
segmental allotetraploids. Lorz (47) is of the opinion that thei emphi-
diploid between these tw~o species would always exhibit the colored
flowers, color-ed seed coat and vining habit of P. eolystachyus if no
crossing-over occurs between interspecific genomes. The expression of
such recessive characters as white flowers, unp~igmented; seed coait, and
bush habit as described by Stall (64) clearly indicates that an exchange
of chromosomes or chromosome segments has occurred between P. lunatus
and P. polystachyus genomes. Such intergenomic synapsis resulted in
heterozygosity at these loci on P. polystachyus chrom~osomes. Subsequent
segregation revealed the recessive characters from P. lunatus.
Several abnormalities were noted in the melotic figures and
products of the advanced generations of amphidiploids. To have been
selected for 12 to 15 or more generations, the amphidiploid plants
exhibited an abnormally large percentage of non-viable pollen. When
stained with acetocarmine it was noted that 20 to 35 percent of the
pollen appeared to be devoid of contents, did not take stain, and
appeared to be smaller than stainable pollen.
Meiosis was found to be irregular in all amphidiploid plants.
Table 3 summarizes the number of normal and abnormal meiotic figures
observed in pollen mother cells from the amphidiploid plants. To be
Plant Pl MI Al TI MI l AllI TIl To tal % Ab no rmal1
Normal 12 25 31 3 71
Abnormal 5 3 8 16
Normal 18 26 4 48
Abnormal 5 16
Normal 2 14 32 10 58
Abnormal 3 8 1 12
Normal 2 17 18 7 44
Abnormal 14 1 2 8
Normal 131 10 13 55
Abnormal 10 3 13
Normal 4 25 10 24 63
Abnormal 8 3 10 21
Normal 37 7 12 56
Abnormal 6 3 9
Normal 60 14 20 94
Abnormal 10 2 8 20
Number of normal and abnormal meiotic figures from
amphidiploid lima bean plants after several
generations of self ing and selection
Me i tic Stage--l
Plant Pl MI AI TI MIlI AI I TI l To tal1 % Abno rmal1
Normal 16 25 13 54
Abnormal 3 3 2 8
Normal 1 3 13 47 31 95
Abnormal 2 2 9 2 15
Normal 2 9 38 8 57
Abnormal 2 6 2 10
Normal 4 20 16 10 50
Abnormal 3 3 1 7
"Pl-Prophase 1, MI-Metaphase 1, Al-Anaphase 1, TI-Telophase 1,
MII-Metaphase 11, All-Anaphase 11, TIl-Telophase 11.
Table 3 (Continued)
sure, not all cells in meiosis could be scored as normal or abnormal
due to the clumping phenomenon (16) particularly at metaphase I and
II. Approximately 10 percent of the cells observed could not be scored.
Abnormalities ranged from 11.1 percent in plants 3 and 12 to 25.0
percent in plant 9. When these figures are compared with the subjective
productivity score based on pod set in Table I it should be noted that
plants 3 and 12 are scored "medium" while plant 9 is scored "low~."
Selecting the most productive plants based on the amount of pod set
may be one method of selecting away from melotic abnormalities in lima
beans comparable to results reported by Kostoff (39) for Nicotiana and
Bremer and Bremer-Reinders (8), Roseweir and R~ees (56), and Hilpert
(27) for rye.
Figures 18 to 24 show some of the laggards encountered in this
material. Laggards are generally present in all melotic stages. As
many as four chromosomes or groups of chromosomes have been found as
shown in Figure 24. Most of these aberrant laggards appear too large
to be single chromosomes. Univalents shown by Fozdar (19) appear much
smaller than these bodies. These laggards resemble the clumped lagging
chromosomes observed by Stall (64). He speculates that these are
clumped P. polystachyus chromosomes which had no synaptic partner in
meiosis. The plant with which he worked was a tetraploid, presumed to
have had three genomes from P. lunatus and one genome from P. poly-
Laggards in the lima bean material may arise from a variety of
sources. By secondary association, a univalent of either P. lunatus
or P. polystachylus may synapse with bivalents of their interspecific
homologue. If the affinity to form further secondary associations is
Figure ~ ~ ~ ~ *; 18--bve h~o icrgaho olnmte elo
plant1 3 shwn ogrsa mtpae1(00)
beoa nepeie aealcd rwn
Figure 19.--Above, photomicrograph of a polle~n mother cell of plant
2 showing laggards at telophase I(1050X); below, an
Interpr-etive camera lucida drawing (2000X).
Figure 20.--Above, photomicrograph of a pollen mother cell of plant
660 showing laggards at metaphase 11 (1050X;); below, an
interpretive camera lucid drawing (2000X).
Figure 21.--Abve, photomicro rap of a olnmte elo ln
2 ho in lagr d attlpae 15X ; eo ,a
inerreiv caealcd raig(00)
Figre 2.-Aboe, hotmicogrph f apolen othr cll f pan
66 sown lggrs t eopas I(05X) elwa
inerre iv camera: lucd dra in (00X)
C - -
~.~e~ .~ ~
Figure 23.--Above, photomicrograph of a pollen mother cell of plant
6 showing laggards at m;etaph-ase 11 (1050X); belowJ, an~
interpretive camera lucid drawing (2000X>r).
~ ' ~
h' i' '
Figure 24.--A2bove, photomicrograph of a pollen mother cell of plant
660 showing loggards at telophase II (1050X); below, an
interpretive camera lucid drawing (20030X).
weak, this would leave a univalent unpaired in meiosis and would be a
source of laggard chromosomes.
The )aggards may result from heterogenetic pairing between inter-
specific homoeologues to form a bivalent. Thus the two complementary
homoeologues may be separated by distance or other barriers at the time
of synapsis, resulting in two coiled, but unpaired interspecific
Another possible source of laggards is the presence of a paracentric
inversion in which a crossover occurs within the inverted segment (49,
67). This gives rise to an acentric fragment which will be lost due to
the absence of a centromere, and a decentric chromosome which will form
a chromatinn bridge at anaphase I due to two centromeres. The bridge
may be broken by tension of anaphasic movement or may remain suspended
between the two nuclei and eventually be lost to both.
If, in fact, an inversion is present in the amphidiploid lima bean
lines, it should not be too surprising that the classical inversion
loop at melotic pachytene has not been observed. The small size of
the chromosomes and the difficulty of locating a pollen mother cell at
pachytene greatly decreases the chance of locating such a loop.
Many laggards observed in this material may be the result of early
disjunction. When synapsis lapses at diplotene, some homologous
chromosomes (and more especially homoeologous chromosomes) may be paired
so loosely that early disjunction may be prevalent. This factor would
surely bias the observations of normal and abnormal meiosis if the
chromosomes in early disjunction were scored as laggards which had
never entered~into synapsis.
Early disjunction was observed by Fozdar (19) even in diploid P.
lunatus. All indications are that the chromosomes which exhibited early
disjunction functioned normally otherwise and were incorporated into
viable gametes. This is apparently not the case in the amphidiploids
Figures 25 to 32 record more of the abnormalities found in this
material and perhaps the source of the laggards. Chromosome bridges
and fragments were reported by Fozdar (19) in the FI of the original
cross but none were reported in the amphidiploid progeny. Stall (64)
did not observe chromatin bridges.
Figure 25 represents what is thought to be very early anaphase 1,
just as the -two groups of chromosomes are beginning to separate. A
chromosome or clump of chromosomes appears to be stretched or suspended
between the two groups, as if it were going to be pulled in both direc-
tions and form the typical chromatin bridge. No acentric fragment
could be seen in this particular cell.
Figures 26 and 27 indicate major disfunctions at anaphase I in
these cells. Several chromatin strands can be seen stretched between
the two groups of separated chromosomes. In some cases large blocks
of chromatin are seen in the area of the original metaphase plate.
These may represent whole chromosomes. In other instances the chromatin
strands are very thin. Some strands appear to transverse the entire
distance across the gap while others appear to stop at some point between
the two nuclei but appear to be under some tension. This may be due to
a chromatin strand so thin that it cannot be resolved with the oil
Figures 28 to 32 further illustrate the presence of chromosome
bridges. The cells may be slightly further advanced in anaphase I than
Fiue2.-bv, htmcorp o o lnmtercl i ln
660shoinga sortchrmain ride a ealy napas
1~~~~~~si- (15Xfblw nitrrtv aealcd rwn
Figure~ ~ ~~~C 26-Aoe htmcorp f olnmte elo ln
2 shwin seera chomain rides t aaphse
(1050%);C" beow anitrrtiec r ucd rwn
~,~ -i. 4~ -
i :ii i
:_~ I :i
Figue 2.--Aove phoomirogrph f a olln mohercellof lan
2~ ~ shwn seej hoai rde taahs
(15X) elw n nerrtvecmealciadrwn
ii~ ii f
Figure: 28.--A'oove, photomnicrograph~ of a pollen mother cell of plant
2 sh~owring a single bridge at late anaphase 1 (1050X);
below~, an inter~pretive camecra lucida drawing (2000X).
'r -- "-
I h ,
Figure 29.--Abo~ve, photomlicrograph of a pollen miothier cell of plant
2 showing a chr-omiatin bridge and chrom~osome fragment at
telophase i (1050X); below, an interpretive camera
lucida drawing (2000X).
P ~L 4
Figure 30.-Above, photomicrogr~aph of a pollen mother cell of plant
2 sho~i ng a chromo 1tin br idge and twJo chromosome f ragmen ts
at telophase I (1050X); below, an interpretive camera
lucid drawing (2000X).
Fiue3.-boe htmcrgaho ole ohrcelo ln
show-ingacrmtnbig it orcrmsm rget
at~ teop s I- (150) beow anitepetv mr
Figure 32.--Above, ph~otonicr-ograph of a pollen mother cell of plant 2
showing twIo chromostin bridges at telophase 1 (1050X);
below, an interpretive: cancroT lucida drawings (2000X).
;;" T111~ :J3i~"~"S~:
those shown in Figures 26 and 27. In this case the numerous small
chromatin strands will have broken and only a larger-, more persistent
bridge remains. In the cells with only one apparent bridge, the
chromaitini is stretched thin either in the middle of two chromosomes
(Figure 28) or at the ends (Figure 30). In either case there appears
to be e tension on the bridge as if a chromatin strand not visible in
the microscope is still connecting the two nuclei.
The possibility also remains that there were only one or twro
major bridges In the cells shown in Figures 28 through 32. In Figure
28 no acentric fragments are apparent and no evidnce Is seen of the
remains of other bridges. Figures 29, 30 and 31 show a single bridge
with one, tw~o, and four fragments, respectively, while Figure 32 shows
what appear to be two bridges.
FollowJing the sequence of events a step further, the presence of
acentric chromatin fragments and clumps of laggards should indicate
the presence of irregularities in pollen grain formation.
Figure 33 shows a tetrad of four normal appearing pollen grains
plus a fifth, smaller pollen grain. Figures 34 and 35 show tetrads
having four large pollen grains with two and three smaller pollen
grains, respectively. These were the result of micronuclei and lag-
gards being excluded from the daughter nuclei at telophase It. As the
pollen grains matured, the cytoplasmic contents of these deficient
pollen grains degenerated. At the time of anthesis these pollen
grains appeared empty and did not stain as did the normal sized pollen
Pollen sterility was found to be associated with meiotic irreg-
ularities by Clausen and Goodspeed (ll), Karpechenko (37), Newton and
Figure 33.--Photomicrograph~of a tetrad showing four
large pollen grains and one small pollen
Figure 34.--Photomicrograph of a tetrad showing four
large and twro small pollen grains (500%).
Figure 35.--Photomicrograph of a tetrad showing four
large and three small pollen grai ns (500X).
Pellew (52), Kostoff (39), Sparrow et al. (63), Bremer and Bremer-
Reinders (8), Jones and Kobayashi (33), Roseweir and Recs (56), Hilpert
(27), Fozdar (19), Oh-aliwal e~t al_. (16), and Stall (64).
Table I: gives a summary of the number of aborted grains per 100
.pollen grains counted. The greatest amount of aborted pollen was
found in the LLLP plant and its F2 progeny. This might be expected
since the, three lunatus genomes should tend to form trivalents while
the polystachyus genome would tend to be expressed as non-synaptic
univalents and be eliminated. These plants produced but few seed.
Of the amphidiploid plants, numbers 660 and 9 exhibited the
greatest average percentage of'sterile pollen, 36 and 34 percent,
respectively. The most pollen-fertile of the amphidiploids were
plants 11 and 12 with 22.4 and 21.8 percent sterile pollen, respective-
ly. Plant 12 was also scored "low" in abnormal meiotic figures and
"medium" in seed pod production. The diploid, 'Fordhook', had 6.0
percent non-stained pollen.
A decrease in the percentage of aborted pollen, in addition to
improved seed pod production, may be a valuable tool for a plant
breeder to use for selection of amphidiploid lima bean lines exhibiting
regular meiosis. It would appear from cytological evidence that plant
2, based on percentage of abnormal melotic cells, and plants 11 and 12
based on percentage of aborted pollen grains, would be the most de-
sirable selections of all the amphidiploids used in this study. In
the final analysis, however, one should keep in mind the more important
goals of a breeding program, and not confine oneself to selecting
plants which are perfect in all cytogenetic respects while ignoring
the practical aspects and economic value of genetic uniformity and
Plant Number I 2 3 4 5 Total Average
'Fordhook' 11 12 3 2 2 30 6.0
2 24 19 20 21 17 101 20.2
3 31 36 29 23 25 144 28.8
4 22 35 17 18 28 120 24.0
5 30 24 32 24 35 145 29.0
6 26 28 21 31 19 125 25.0
9 30 26 40 35 39 170 34.0
10 20 29 26 36 19 130 26.0
11 28 27 14 26 17 112 22.4
12 26 16 22 16 29 109 21.8
660 17 18 13 66 66 180 36.0
986 19 29 12 25 34 119 23.8
985 28 32 24 38 19 141 28.2
LLLP 57 57 58 75 70 317 63.4
LLLP F2-T 61 62 73 62 56 314 62.8
Number of aborted pollen grains in 100
pollen grains counted from each
of 5 flowers fromt diploid
and amphidiploid plants
70 50 46 69 53
The selection of more fertile types should enhance the possibility
of obtaining a backcross to the diploid P. lunatus. The triploid
thus obtained should have two lunatus genomes and one polvsta~chys
genome. Its fertility, i.e., the production of viable pollen carrying
polystachyus chr-omosomes, should be dependent on the synaptic ability
of the two lunatus genomes.
Tables 5 and 6, and Appendix Tables 8 through 21 give the pollen
grain diameters of diploid 'Fordhook' and all amphidiploid plants
used in this study.
Table 7 shows the analysis of variance table of the amphidiploid
plants. 'Fordhook' was not included in the analysis since Fozdar
(19) showed the pollen diameter of 'Fordhoo'k' to be significantly
smaller than the amphidiploid. The "F" test (62) indicates a sig-
nificant difference between the means at the .05 level of significance,
Figures 36 through 51 give the frequency distribution curves of
the pollen grain measurements of the amphidiploid plants as well as
diploid 'Fordhook'. Several of these graphs indicate a shift to a
more uniform pollen grain size while others show a greater variation
in the size distribution. Most size distribution figures presented
by Fozdar (19) have a greater range and generally a flatter curve
than many of the plants used in this study. The trend appears to be
toward a sharper curve such as is seen in Figure 37 or in Figure 39.
Changes in pollen grain size may be caused by changes in the
chromosome numbers. The change in ploidy level itself brings about
a proportional change in pollen grain volume. A change from diploidy
to tetraploidy causes a doubling in volume. An aneuploid loss of
chromosomes can also effect pollen grain size. Even as small a loss
Number 12 3 it 5
Pollen grain diameters in microns
from 5 flowers of plant 3
Number I 2 3 4 5
_ _~ ____
Pollen grain diameters in microns
from 5 flowers of plant 5
54. 00 -52.65
Source of Variation d/f S. Squares M. Square F
Buds (b-1) 4 1,031.97
Plants (p-1) 14 3,329.22 237.80 2,05"
E rror ( a) (b1(-)56 6~,505.70 116. 17
Plots of plants (74) 10~,866.89
Buds (b-1 (4) (1,031.97)
Observations (o-1) 19 1,091.73 57.46 1.49
Error ~(b) (b1(-)76 2,9 24. 13 38.47
Plots of observations 99 4,015.86
Buds (b-1) (4) 1,031.97
Plants x observations (p1(-)266 13,490.86 50.72 1.75"
Error (c) b1(-)o1 1,064 30,7 20.09 28.87
Total boo-1 1,499 59,093.70
Analysis of variance of pollen grain diameters of advanced
generations of amphidiploids derived from the cross
P. lunatus var. 'Fordhook' x P. polystach~yus
Sinfcata 05 lv l
" Significant at .01 level.