in 17 Osmia-spccies of the aduncn-Group- 163 



Before discussing the separate species, it may be worth 

 while to describe in some detail the general structure of 

 the 6th ventral segment in the achmca-Group and the 

 general nature of the specific characters which occur in it. 

 To draw it undamaged from its retirement under the 

 5th ventral is not always easy. My own method is, after 

 relaxing a specimen, to force apart with a dissecting-needle 

 the 5th and 6th dorsal segments. Being rather firmly 

 attached to the base of the latter, the 6th ventral generally 

 comes out with it. It is then seen to be formed of several 

 distinct layers superposed one upon another, some quite 

 thin and hyaline, others more substantial and darker. 

 Most of these at least do not extend to the base and apex 

 of the segment, but occupy a part only of its full length. 

 The actual base is pretty solid. It is deeply excised, 

 accordingly bidentate — the two teeth are attached externally 

 to the 6th dorsal by a membrane, which must be cut 

 through carefully, if the segment is to be extracted entire. 

 Beyond the basal excision begins the thickest and most 

 substantial part of the segment. We see here, first, but 

 (owing to their transparency) only in certain lights, two 

 adjacent flakes of thin white membrane, attached only at 

 their bases (the rounded apices being quite free) to the 

 underlying layers of substance. Below these flakes, and 

 partly at least projecting beyond them apically, is a much 

 more solid transverse layer (or conglomerate of layers) 

 divided longitudinally into two well-marked lobes — dark, 

 punctured, and more or less pilose, especially towards their 

 apices laterally. These I shall call in the following de- 

 scriptions the " main lobes " of the segment. From between 

 these lobes, at a rather lower level, originates the apical 

 membranous appendage which I shall call the " process." 

 It, also, usually assumes a somewhat bilobed form ; but in 

 two species it is, instead, terminated by a single central 

 (spine-like) prolongation. The base of the " process " 

 rarely occupies the whole space between the converging 

 margins of the " main lobes." More usually it has a con- 

 stricted petiole-like base, from which the lobes of the bifid 

 apex branch out more or less in the lateral direction, 

 making the process as a whole roughly Y-shaped in some 

 cases, T-shaped in others. The petiole of a Y-shaped 

 process is mostly long and narrow, that of a T-shaped much 

 more transverse (compare Fig. 7 with Fig. 11). Round 

 these two types, the Y-shaped (ctementaria) , and the 



TRANS. ENT. SOC. LOND. 1901. — PART II. (JULY) 12 



