SeterogynU paradoxa. 7l9 



flew to a neighboaring twig, crawled up it, and settled 

 down in the same way, but if really frightened, they would 

 soar for a minute or two, and if there was any wind were 

 liable to be carried right away. Only one or two were 

 seen, early in the day, that were flying, without having 

 apparently been disturbed. I took some more the follow- 

 ing day, but a few days after had difficulty in finding one 

 or two; it was in fact practically over, although a larva or 

 two (ichneumoned ?) could still be seen. Its period of 

 flight is therefore (in a late season) from about June 

 20th — 30th. To avoid circumlocution I propose to call 

 tiiis form of H. paradoxa — var, Bcjarcnsis. 



On July 3rd, on the Sierra de Bejar, above Candelario, 

 at a height of perhaps 5000 feet, I took several larvas of a 

 Hctcrogynis on Cytisus lyiirgans, which grew from about 

 that elevation up to fully 6500 or thereabouts — a species 

 that clothed large areas of the Sierra, and made a mass of 

 most splendid bloom at about 6000 feet, but below was 

 already out of flower and had apparently bloomed there 

 but sparsely. I took two or three more larvoe later. On 

 July 12th these larvae produced two % imagines. So on 

 the 16th I visited the locality to see if I could not capture 

 some males. Sembling was of no avail, but I succeeded 

 in capturing two or three dozen specimens. Only one or 

 two of these were taken by observing them at rest, nearly 

 all being disturbed by walking through and beating the 

 broQm bushes. The proportion of worn specimens showed 

 they were a day or two past their best. This variety I 

 call Canddariiv. 



Later still, July 18th, at Piedrahita, some twenty-five 

 or thirty miles from Bejar, and on an outlying ridge of the 

 Sierra de Gredos, separated from the Bejar range by a 

 rather wide valley, I met with a third form of Hctcrogynis. 

 It was then only just coming oift at its lowest levels, at its 

 highest, 600 to 1000 feet higher up, only larvae were to 

 be seen. This form {Picdrahitm) is therefore at least ten 

 days later than Candclari/v in the date of its appearance. 

 The chief interest attached to these three groups or 

 colonies of H. paradoxa occurring within a limited region 

 at successive dates, lies in the circumstance that each 

 colony presented very definite characters by which it was 

 distinguishable from the others. 



Each group is thus probably entitled to a separate name 

 on its merits ; for convenience in speaking of them, how- 



