THE POLLINATION OF SATYRIUM BICALLOSUM, THUNB. 149 
each spike bears many flowers, and the plant itself grows in considerable 
abundance on the open grassy flats, almost every flower that I have examined 
has been fertilised.” 
Bolus 1 has made observations upon the frequency of the removal of the 
pollinia in Disperis capensis and D. villosa, possibly by night-moths. He 
observed no insect visits. 
My attention was first directed to Satyrium bicallosum, Th. by a remark 
made by Dr Bolus 2 . He says: “In the structure of the column this species 
differs from any other known to me. The anther, instead of hanging vertically 
under the rostellum with its glands pointing forward, as is the typical structure 
in the genus, appears as if pushed up against the apex of the column, so as to 
lie nearly horizontal, with the glands turned to the back of the flower, and 
the stigma completely covered and hidden from view. The economy of its 
pollination is a puzzle to me, for the appearance is as if the process were 
carefully prevented.... But I have not had the opportunity of watching it, 
and have never seen any insect or trace of insect-work on the plant.” These 
remarks led me to suppose that possibly this was a case in which self-pollina- 
tion occurs, particularly as the flowers are rather small and inconspicuous, 
although belonging to a genus with typically large and brightly coloured 
flowers. Through the kindness of Miss Watermeyer, of Jonkers Hoek, Stellen- 
bosch, I was able to secure a large number of inflorescences for examination, 
and also to make observations upon the plants growing under natural con- 
ditions. The first group of inflorescences examined showed evidence of insect 
visits. Quite regularly, the lower flowers on a spike were found to have both 
pollinia removed, whereas open flowers in the upper part of the spike had only 
one pollinium extracted. This observation at once indicated that cross-pollina- 
tion might occur, and that a search would have to be made for the stigmatic 
surface, which Dr Bolus supposed to be completely hidden from view. As the 
account and figures of the column of this flower were made by Dr Bolus foi 
purposes of systematic work, the results of a more detailed examination of 
the structure must now be given. As the structure of the column is intimately 
correlated with that of the perianth and bract, the latter will be first described. 
The inflorescence is a spike, from 3 to 10 cms. long, with the flowers arranged 
in a crowded spiral, and each flower in the axil of a large, curved bract. The 
acuminate apex of the bract reaches to the galeate posterior petal, and, as the 
blunt tip of the petal (labellum) is somewhat depressed, each flower has two 
lateral entrances (Plate vi, Fig. 12). This great depression of the galea, 
forming two eye-like openings, together with short bracts in the upper part 
of the spike, constitutes the variety ocellatum Bolus 3 . It is obvious, o\\ in_ 
to this peculiarity of structure, that only one pollinium (the light hand oi t ie 
left hand one) can be removed at once, thus accounting fox the numerous ca 
1 Bolus, 1918, p. 42. 2 Id. p- xxii. 2 Id- P- 4L 
