( xe ) 
in cases of the same kind which occurred to myself when 
collecting.” 
Prof. Poulton remarks: “ It is tantalising to reflect upon 
the number of interesting and important questions which 
could be now decided if” the practice of collecting and care- 
fully labelling all specimens captured in coitéd “had prevailed 
during the past fifty years.” 
I now bring forward a few facts bearing on the question 
of the preferential mating of varieties, not because it is con- 
sidered that sufficient, evidence has been produced to prove 
anything, but rather with a view of directing more attention 
to this important point. On Bugalla Island, L. Victoria, 
during 1912, I found that a small and extremely variable 
Syntomid moth, Lpitoxis albicincta, Hmpsn., was very abun- 
dant; I had previously met with this species at Jinja in Usoga, 
on the mainland, and at Mpumu in Chagwe, but not in such 
abundance as on the island. The moth varies equally in 
either sex, from a form all black without markings to a form 
with large and well-defined white spots. The first specimen 
I took was one of the fully spotted forms, on a grass stem 
(for it is always found in open grassy places) on Mpumu Hill, 
Chagwe. It hung feet upwards from a bending stem, and, 
when approached, held its wings in a peculiarly twisted way, 
hanging perpendicularly from its back but with the ventral 
surface facing forwards. I have obtained altogether 58 
specimens in various localities, of which 22 were actually 
in coitd. These can be arranged under six types, but of 
course no one type can be quite sharply marked off from 
another. 
Before making any remarks upon the specimens grouped as 
on p. xci, it must be said that they were not collected alto- 
gether as they came. It will be seen, for instance, that Type VI 
comes mostly from the mainland; and although my impres- 
sion is that it was relatively more abundant on the mainland, 
more material is required. For when I got on to the island 
I remembered that I had already caught Type VI on the 
mainland, and devoted myself rather to getting a good series 
of varieties than to determining the relative frequency of 
each. It is also suggestive, that Type I does not seem to 
