December, '11] cushman: hosts and parasites of brucuid^ 491 



attacked by B. compressicornis Schaef. This plant is one of the most 

 abundant weeds along ditchbanks, headlands, and other waste lands 

 throughout the region. 



In regions where the parasites have constant access to their natural 

 hosts, they only occasionally leave these to attack another insect. 

 However, if, through some catastrophe, they are deprived of their 

 normal hosts, they search for and oviposit in other suitable hosts. 

 An instance of this catastrophic transfer of parasites from their normal 

 host to another was furnished in the summer of 1907 at Victoria, Texas. 

 During that season the huisache trees over a large area produced almost 

 no pods, the normal breeding place of Bruchus sallcei Shp., due to the 

 activities of a small weevil {Tychius sp.), which breeds in the flowers. 

 A remarkable increase in the percentage of parasitism of the boll weevil 

 was traced to this occurrence by the breeding from infested cotton 

 squares of large numbers of Cerambycohius cushmani Cwfd., normally 

 parasitic on B. sallcei. 



Of the other parasites known to attack both the boll weevil and 

 bruchids, the species of Eurytoma have not been determined; Larioph- 

 agus texanus Cwfd., though attacking certain of the bruchids abun- 

 dantly, has but seldom been recorded from the boll weevil; Br aeon 

 mellitor Ashm., Microdontomerus anthonomi Cwfd., and Habrocytus 

 piercei Cwfd., all normally weevil parasites, have been only occasion- 

 ally reared from bruchids. 



Certain of the parasites seem to be confined to the bruchids as hosts. 

 In this class are Urosigalphus bruchi Cwfd., Glyptocolastes bruchivorus 

 Cwfd., Heterospilus bruchi Vier., and H. prosopidis Vier. The species 

 of Horismenus, which have not been determined, are even more 

 closely restricted, some being apparently confined to a single host 

 species. 



If the author seems to have ignored the practice of the European 

 authors of giving the family name of this group of beetles as Lariidse 

 instead of Bruchida?, and of dividing the genus Bruchus, he must plead 

 a lack of the systematic knowledge of the group necessary for ])lacing 

 the different species in their proper genera. To this circumstance 

 also must be attributed the lack of any attempt at a logical arrange- 

 ment of the species considered. The information given is simply 

 contributed as an addition to the biological knowledge of the group. 



The thanks of the writer are due and are herewith extended to Mr. 

 Charles Schaeffer, who determined most of the Bruchida^ and to 

 Messrs. J. C. Crawford and H. L. Viereck, who determined the 

 parasites. 



Bruchus julianus Horn, Proc. Cal. Acad. Sci., Ser. 2, V, p. 410. 



This species, the largest of the genus in the United States, was 



