Native Hawaiian Cotton — S tephens 
393 
on the outer surface of the calyx alternating with 
the bracteoles, and there is a leaf nectary on the 
midrib of the under surface of each leaf. It is 
thus possible for insects to obtain nectar without 
entering the flowers at all. These characteristics 
do not suggest floral adaptation to any specific 
insect vector (Grant, 1950:392). 
In tomentosum extrafloral nectaries and leaf 
nectaries are lacking entirely, but a true floral 
nectary can be seen in vertical section. There 
is reason to suppose that the latter is nonfunc- 
tional or only weakly functional. When the 
corolla is removed from a fresh flower of bar- 
badense or hirsutum, a pool of nectar can be 
seen within the calyx cup, but the secretion is 
weak or absent in tomentosum. This observation, 
based on the relatively few flowers available at 
the time of study, was confirmed later by Mr. 
Frank Johnson. He informs me that only a 
fraction of the flowers he examined appeared to 
contain nectar. It seems likely, therefore, that 
insects would visit tomentosum flowers primarily 
to feed on pollen or to gather pollen. Further, 
because a limited number of plants are in flower 
at any one time in tomentosum populations, they 
would not be likely to furnish a main source of 
supply to pollen-gathering insects. Under these 
circumstances a mainly autogamous breeding 
system is to be expected, with visits by pollen- 
gathering insects restricted to temporary periods 
of over-all pollen shortage. 
In Central America and the Caribbean, the 
relatively small and scattered populations of wild 
and dooryard forms of G. hirsutum do not im- 
press the observer as being heavily stocked with 
pollinating insects. Those commonly found in 
the flowers are small ants and beetles which 
probably function more effectively as agents of 
self- rather than cross-pollination. The situation 
in tomentosum populations appears to be similar, 
though the number of observations which could 
be made was too small to be very reliable. 
Unusually heavy and persistent rains occurred 
throughout March and well into April in this 
normally dry habitat; thereafter there was a 
temporary cessation of flowering until toward 
the end of May. In the latter part of April 
the combined results of three separate searches 
through the Oahu populations yielded less than 
50 flowers. It was noted that although honeybees 
and small moths (occasionally) and carpenter 
bees and scavenger flies (rarely) were visiting 
Sida flowers in the immediate neighborhood, 
none of these potential pollinating agents visited 
tomentosum flowers. Most of the latter were 
entirely devoid of insects, though occasional 
flowers containing a few small ants or fruit-bud 
beetles ( Conotelus mexicanus) were seen. 
These observations, though limited, do not 
suggest that insects play an important role in 
the breeding system of tomentosum. On the 
other hand, a strictly autogamous system is dif- 
ficult to reconcile with the fact that tomentosum 
flowers, like those of other wild forms of the 
New World species, have long styles with the 
receptive stigmatic surfaces borne high above 
the staminal column (cultivated forms usually 
have short styles with the stigmatic surfaces in 
contact with, or immediately above, the staminal 
column). An interesting feature of the tomen- 
tosum style is the fact that it is often recurved, 
assuming a crozier-like structure with the stig- 
matic surface nearly approximated to the stami- 
nal column. 
2. The Effects of Insect Introduction 
In oceanic islands like the Hawaiian group 
with a numerically poor indigenous insect fauna 
(Zimmerman, 1948:94-95), the introduction 
of beekeeping as a local industry could have 
potentially disturbing effects on the taxonomy of 
the local flora. A likely example is provided by 
the dooryard cottons ( G . barbadense) , which 
still may be found in gardens and along road- 
sides on the islands. As early as 1812, cotton 
was introduced into Oahu (Jones, 1937:17) 
and, later, in the early 20th century, serious 
attempts were made to establish it as a crop 
(Krauss, 1909:1-16). Commercial varieties of 
Sea Island ( barbadense ) and Upland ( hirsutum ) 
were both introduced without lasting success. 
Around 1857 the first shipment of honeybees 
arrived from California (Eckert, 1951:1). Ac- 
cording to Van Dine and Thompson ( 1908:8),, 
Prosopis soon became the primary source of 
nectar, and on the western side of the island 
apiaries were moved to the coastal areas under 
the shelter of the Prosopis forests, where many 
of them are found today. During the past 10 
years this source of nectar has been seriously 
diminished by the activities of an immigrant 
moth, Ithome concolorella , which destroys the 
