95 



selection of individuals able to escape the notice of the parasites and 

 so modify the race. But such modification may be very slow; for exam- 

 ple, the Prosopophora on the before-mentioned Atriplex is very little 

 different from a Japanese species from which it must have parted com- 

 ])any long ages ago. On tlie whole, selection by i)arasites is probably 

 a less imi:)ortant factor in coccid evolution than one had been ready to 

 assume. 



In our ordinary accounts of coccid enemies we omit some factors 

 which may possess imi)ortaiice. ]Mr. Xewstead has lately given some 

 interesting observations on the feeding of birds on coccids. Mr. Bar- 

 rett, above cited, found that the San Jose scale disappeared from his 

 Walker's Seedling i)ear trees, which were very badly infested, when 

 Bryohia pratensi.s ai^peared upon the scene. I saw the trees with the 

 Bryobia upon them, and no scale to speak of. Mr. Steel, also of Las 

 Cruces, X. 31ex.. had some peach trees badh' infested by San Jose scale. 

 When I examined the trees this year there was no living scale, and the 

 trees were swarming with a small ant. which Mr. Pergande determines 

 as Tainnoma (uiale.^ For two years these trees had borne no fruit: now 

 they are fruiting. 



These rough notes may serve to indicate some of the lines of work 

 desirable in the future, if we are to understand our CoccidiC. Xo greater 

 mistake could be made than to sujopose that I or any other person can 

 from cabinet specimens alone settle everything. We must have obser- 

 vations on the living insects. 



HOW SHALL WE IMPROVE OUR COLLECTIONS? 



By C. P. Gillette. Fort Collnts, Coin. 



A correctl}' determined collection of insects is about as important for 

 the best entomological work as a good library. The writer, in receiving 

 answers from various college and station entomologists in regard to 

 their collections, has been surprised to tind that there are so few really 

 good collections and that there are so many institutions that have 

 practically no collection of determined insects at all. This is not as it 

 should be, and with a little effort of the right sort on the part of the 

 entomologists in charge 1 believe it need not long be so. 



Collections may be made in various ways. Some of the older insti- 

 tutions could well afford to purchase good collections, but few of us 

 can depend upon any considerable sums of monej' for this purpose, and 

 it is probably about as well that we can not. What I wish to urge at 



* First described by Ernest Andre from a specinieii collected by the present writer 

 in Mexico in 1893. In Proc. Cal. Ac. Sci.. May. 1804. pa.uo l»i3, Mr. Perirando considers 

 it a variety oi' T. pruinosum Ro«r. and records it Irum California and Lower Cali- 

 fornia. It is new to Xew Mexico. 



