POLYOMMATUS ICARUS. 215 



reported by Sparre- Schneider from June 7th to August 15th in 1891 ; 

 in 1890 it was not observed earlier than June 11th, and in 1892 as 

 late as August 17th, but in his opinion there was only a single brood 

 in each case, though he found a second brood as far north as 

 Christiania (slightly north of the Orkneys) in August 1876. Bang- 

 Haas reports that it is common in Denmark in May and June, and 

 again in August and September, thus clearly indicating two broods. 

 It is reported by Assmuss to be single- brooded in Russia in the 

 Moscow, Kaluga, and Tambov Governments, though found from June 

 to the end of August; in the Government of Viatka, where it is found 

 from the end of May to August, Kroulikowsky reports it as generally 

 single-brooded, though- there is sometimes a supplementary second 

 brood consisting of very small specimens ; in the Governments of 

 Kasan and Vologda where he reports it as occurring only in June and 

 July (in the latter case he specifies early July) it is of course only 

 single-brooded. Nolcken, on the other hand, says that it occurs in 

 the Baltic Provinces (Esthonia, Livonia, and Courland) in two broods, 

 from the end of May far into June, and from the end of July to 

 September, and is not scarce. An unaccountable case of apparent 

 single-broodedness is reported by C. Seymour Brown from the Island 

 of Capri, where he has only taken this species in the early summer, 

 though elsewhere at this latitude there are at least three broods, and 

 the species seem specially to flourish in islands. Frey reports it as 

 occurring in Switzerland in two broods up to 7000 ft., w T ith an 

 occasional third autumn brood in September in favourable years 

 and localities, whereas Favre states that it occurs in the Valais 

 from April to October. Wheeler is of opinion that in the plains and 

 the subalpine region there are always three broods, but only two at 

 the higher levels. Indeed, in all places where the first brood appears 

 as early as the first half of April, a third brood may be confidently 

 expected. At Brindisi, for instance, there are at least three broods, as 

 the following notes from Simes (in I'M.) serve to show: — March 28th, 

 1905; March 27th, 1906 ; Mav 9tb, 1910; May 23rd, 1905 ; Jane 

 19th, 19U6 ; July 8th, 1905 ; Angst 21st, 1906. In localities still 

 further south where the first brood appears early in March, there 

 are at least as many as four broods in the year, as the following- 

 notes on the appearance of this species in Malta in 1902, com- 

 municated (in litt.) by Fletcher will suffice to prove. March 8th, 

 one $ ; March 26th, a few ^ s, worn; April 6th. Js common, 

 quite fresh, only one £ seen ; April 14th, a few, worn ; April 17th, 

 one worn J only ; May 13th, common and quite fresh ; May 15th 

 and 17th, common and fresh ; May 24th and June 2nd, common 

 and fresh, some $ s of the var. celina, all ? s dark ; June 7th, 

 common but worn ; June 14th, abundant, fresh $ s of the var. celina, 

 fresh 2 s of the var. ru/ina ; August 18th, $ s of the var. celina fairly 

 common, but no 5 s seen. From this it would seem that a small first- 

 brood appears early in March, and is worn out in about three weeks, a 

 second emergence, but not necessarily, or indeed probably, a second 

 brood, appearing at the beginning of April ; a genuine second brood, 

 the offspring of the March specimens, appears in greater numbers in 

 Mav, joined later in the month by the off-pring of the April emergence, 

 which show the celina form in the $ s ; a third brood, still more 

 abundant than the second, but of a different form in the £ , begins to 



