260 HYMENOFTEEA. 



fourth joints of the antennae ; the form of the clypeus ; the manner in which the head 

 is developed or narrowed behind the eyes, and also its width compared to the thorax ; 

 the shape of the apex of the pronotum and of the median segment, and also of the 

 median area ; the form of the basal ventral segment of the abdomen, more particularly 

 of the central keel. As regards the form of the keel, it has been used by Sichel and 

 Eadoszkovsky, in their " Essai d'une Monographic des Mutilles de l'ancien Continent " 

 (Horse Soc. Ent. Eoss. vi. pp. 139 et seq.) to classify the males. 



I have only been able to find one record elucidating the natural history of the 

 American Mutillidae. Blake, in his Monograph (Trans. Am. Ent. Soc. 1886, p. 179), 

 simply says that they " are now well known to be parasitic, especially on various species 

 of Bombus" ; and Cresson (Trans. Am. Ent. Soc. 1887, Suppl. p. 108) in practically 

 reproducing this statement very properly points out that they must have some other 

 hosts than humble-bees, inasmuch as they are found in abundance where no Bombi are 

 known to occur. 



That they are parasitic and solitary in their habits may be taken for granted, as also 

 their love for hot sandy localities. The instance alluded to above as regards the habits 

 of one of the American species is given by Westwood (Introduction to the Modern 

 Classification of Insects, ii. p. 214), who remarks that "from information given by 

 Mr. MacLeay to Mr. Kirby, it appears that a large North-American species (Mutilla 

 coccinea* = ' l .occidentalis,lAim.) is very active, taking flies by surprise, probably for 

 the purpose of storing its cells. Its sting is so powerful that a person stung by 

 one lost his senses in five minutes, and was subsequently so ill that his life was 

 despaired of." 



That they must have other hosts than Bombi seems quite clear from the records, 

 scanty as they are, of the habits of European species. One most accurate observer, 

 Giraud (Verh. zool.-bot. Ges. Wien, 1863, p. 1307), relates how he found four cocoons 

 identical in every respect, from which he bred two Ammophila heydenii, Dahlb., and one 

 Mutilla differens, Lep., the deduction being apparently conclusive that the Mutilla is a 

 parasite of the Ammophila. Sichel (cf. Horse Soc. Ent. Ross. vi. p. 11) concludes from 

 his observations that Mutilla incorwpleta, Lep., is in all probability a parasite of a small 

 species of Halictus (a solitary bee), and the same observer (op. cit. p. 12) believes that 

 Mutilla coronata preys on Larra anathema. 



The classification of the Mutillidae is a work of extreme difficulty, and in view of the 

 very large number of species ranged under Mutilla (between 900 and 1000) it would 

 be much more convenient if the genus could be satisfactorily split up. It seems, how- 

 ever, almost impossible to do this in anything like a natural manner, even with the 

 females, while so few of the males being allocated to their respective females further 

 increases the task. Moreover, what might be a satisfactory system of arrangement for 



* This name is not to be found in Smith's British Museum Catalogue, nor in Blake's Lists. 



