( xxiii ) 



down here to warrant the theory of the wet and dry forms of 

 H. bolina, but as soon as I got your last letter I began looking 

 out for a rain-gauge, and have had one going for a few weeks, 

 so in time I shall be able to give you accurate information. 

 At present I should say that the following is a pretty fair 

 description of our climate as regards rain. Jan., Feb., March. 

 April — wet; May — fine; June, July — wet; August — wet, but 

 with fine intervals; Sept., Oct., Nov. — fine; Dec. — wet, but 

 chiefly in the evenings and nights. Rain has fallen on 17 

 of the 25 days of this month, and the total rainfall for these 

 days is 5*565. At the same time the days are beautifully 

 sunny until late in the afternoon, and there are plenty of 

 butterflies about. Thunderstorms are fairly frequent from 

 now till the end of February, and it is in these three months 

 that we get the heaviest rainfalls. In June and July the rain 

 is much more like that in England — not very heavy but 

 continuous, sometimes for a fortnight, with intervals of not 

 more than a few hours at a time. I thought I had answered 

 your question as to the date when I first saw H. bolina. I 

 sent the first specimens I caught to Tring, and the date was 

 marked on the envelope. I should say it was in 1912, though 

 it might have been a little earlier. [The first date is Jan. 

 1911.] For some time I took the butterflies for H. misippm. 

 I think it would be fairly safe to say it was not before 1911 

 that they appeared at Ambinanindrano. The female is still 

 extraordinarily shy. I don't think I have seen more than 

 3 or 4 specimens since I came out this time; whereas the 

 male is always en evidence." 



The specimens were sufficiently numerous and had been 

 observed long enough to make it worth while to tabulate the 

 seasonal forms of the males according to the months. Two 

 examples out of the 82 were omitted because of insufficient 

 data. The seasonal forms of the specimens had been deter- 

 mined from the patterns, especially of the under surface, 

 before Prof. Poulton had received information upon the climate 

 of Ambinanindrano. 



In the following table W stands for " wet season form " ; 

 W — for " nearly full wet '* ; I for " intermediate " ; I— W and 



