HOW COLIAS EDUSA WINtERS. 251 



Dale writes :—" I found a caterpillar of ('. edusa last October at 

 Charmouth ; it changed to a chrysalis on our journey home, and died 

 in the act of emerging at the end of March." It would be absurd I 

 presume to assume that this pupa was not kept more or less under 

 artificial conditions. 



Mr. C. W. Dale, junr., records that " It (the imago) may be found 

 throughout the year on the Continent, even as early as February, at 

 Malta." This is a loose general statement, but we may assume that 

 the appearance of C. edusa at " Malta in February " has been gleaned 

 from a reliable record ; at any rate, C. rdma, is recorded by Mr. T. 

 Blackmore as common on the Dar-al-Clow range of hills 20 miles 

 S.W. of Tangier (/';. .1/. M., vol. v., 299), the record reading, 

 " Common at the end of February on the Dar-al-Clow, a 

 range lying some 20 miles S.W. of Tangier. A few specimens 

 taken close to Tangier a month later." Mr. G. T. Baker 

 records it as common in February at Lambessa, and again in 

 June with var, helice, at Guelma in Algeria {K.M.M., vol. 

 xxii., p. 251) ; and all spring reports from the European and 

 African shores of the Mediterranean report this species as occur- 

 ring regularly in February and March and again in June. Mr. 

 Christy, from eggs laid at the end of August, 1892, discovered feeding 

 larvJB on October 20th. These were removed to a warm greenhouse, 

 where they pupated on December 20th, and after that were kept 

 " warmer than they would have been in summer out of doors." This 

 appears pretty strong evidence that the larvae must have attempted to 

 pass the winter in the larval stage under natural conditions. 



I was for a long time a believer in the assumption (that I suppose 

 I must have assimilated as a very small boy from Newman) that 

 C. rdiisa hybernated in the imago state. Its occurrence in October 

 and November in 1877 (my first real experience of an " cdum year") in- 

 tensified this belief. But my observations on this species during the last 

 two summers at the end of July and throughout August — in Savoy, 

 Piedmont and the Austrian Tyrol — began to shake my faith in my 

 former belief. I found C. cdum appearing very much as it does in 

 England, and that it really was less abundant in the early part of 

 August than it was in its special years in August at home. Now it 

 had always been a pet theory of mine that early August would have 

 found it going over in the southern valleys, and that larvte would be 

 feeding for the September brood. But I can only assume from the fact 

 that from the middle to the end of August is their favourite time, that their 

 progeny do not emerge until October (and then perhaps only partially), 

 so that at least three appearances coincide with those in Britain, June 

 (with our immigrants), August and October. But, as we have seen, 

 we miss entirely the February brood, and get a clue as to the continued 

 extinction of the species here, whilst at the same time the coincident 

 appearances in May, August and October suggest how far our specimens 

 are influenced by " heredity," and how far their habits remain as nearly 

 identical as possible with those of their parents from warmer and 

 more favoured climes. 



We have now to consider, then, how the species passes its time 

 from the October and November emergence until the February emer- 

 gence, a period very little in excess of the time that elapses between 

 the August and October emergences, and we may at once conclude that 



