40 
PACIFIC SCIENCE, Vol. IV, January, 1950 
chromosome number of 18 shows that the 
plant is not a double diploid, and, therefore, 
in view of its full fertility, very probably not 
an interspecific hybrid. A review of the his- 
tory of ftavicarpa points to a more probable 
origin as a mutation of P. edulis. There are 
several notable differences between typical 
edulis and form ftavicarpa, however, includ- 
ing a high degree of cross incompatibility 
The two forms of the species deserve more 
study from the standpoints of both cytology 
and genetics. 
I _i 
Fig. 3. Meiotic metaphase I in P. edulis form 
ftavicarpa showing 9 bivalent chromosomes. 
No flowering material of P. Pfordti, P. 
maliformis X laurifolia, X P. caerulea, or P. 
princeps-coccinea could be obtained by the 
writer for studies of chromosome behavior at 
meiosis in these interspecific hybrids. All 
four appear to be completely sterile. 
P. suherosa is the sole representative of the 
2n = 24 group. Microsporogenesis is normal 
in the species, but there is evidence of strong 
/Oa* 
Fig. 4. Meiotic metaphase I in P. suherosa 
showing 12 bivalent chromosomes. Broken lines 
indicate pairs of bivalents in secondary association. 
secondary association between the bivalents 
at the first metaphase (Fig. 4). Secondary 
association in meiosis has been regarded as 
indicative of an earlier polyploidal origin 
(Darlington 1932: 219-223). 
P. suherosa is also the sole representative 
of the 2n = 36 group. This 36-membered 
form is undoubtedly an autotriploid deriva- 
tive of the 24-membered form, constituting 
a separate chromosomal race within the spe- 
cies. There are no conspicuous morphological 
differences between the two forms. The only 
distinguishable differences are to be found in 
the slightly larger leaves and a slight intensi- 
fication of anthocyanin pigmentation in the 
young stems and on the dorsal surfaces of 
the sepals of the flowers of the triploid form. 
The species, however, is highly polymorphic, 
with all degrees of intergradation among its 
numerous variants, so that even these differ- 
ences are not entirely reliable as a means of 
distinguishing the two races in nature. The 
triploid race appeared spontaneously among 
wild populations of the diploid race, the first 
collection being made in 1937. It produces 
fertile seeds, and has continued to reproduce 
and spread under natural conditions. Exami- 
nations of all herbarium sheets of collections 
made prior to 1937 in the Bernice P. Bishop 
Museum at Honolulu reveal none which 
might conceivably be the triploid form. 
Contrary to what might be expected of an 
autotriploid, triploid P. suherosa goes through 
microsporogenesis with a fairly high degree 
of regularity and produces a preponderance 
of normal quartets of microspores. Mega- 
sporogenesis must be equally little disturbed, 
for the fruit produces a complement of seeds 
more or less comparable to that produced by 
the diploid. Multivalent chromosome con- 
figurations were observed in a few micro- 
sporocytes, indicating that there is some dis- 
turbance to normal bivalent pairing. In an 
occasional sporocyte, all orders of association 
from univalence to sexivalence have been 
observed. Anaphasic separation into equal 
