SCIENTIFIC CULTIVATION OF BRITISH BEES. 147 



More extensive collecting has also shown that some of 

 his species are merely varieties of others, which have 

 thus been brought to their authentic type. This also 

 could only be proved by experience, for it is remarkable 

 how very Protean some species are, whilst others are 

 almost rigidly unchangeable. Evidently there does exist 

 a line of demarcation between distinct species, which 

 only requires to be diligently sought to be found, 

 obscure as it may appear to be, but which the insects 

 themselves obey, for however closely species may some- 

 times approximate, yet I do not believe, as I have before 

 expressed, that they ever permanently coalesce, and that 

 they are always as distinctly separate as are asymptotes. 



As Mr. Kirby's work is in few hands, or perhaps not 

 readily accessible, I will give here a summary outline of 

 it, with the names of the genera with which his families 

 coincide. 



In this work he established only two named genera — 

 Melitta and Apis. 



His genus Melitta, which is equivalent to the subse- 

 quent subfamily Andrenidce, he divides into two sections, 

 * and * *, the first containing two families, a and b, (these 

 we call genera, and they are now named Colletes and 

 Prosopis) ; the second section * * contains three fami- 

 lies, a, b, c, (a, is Sphecodes, b, Halictus, and c comprises 

 our three genera, Andrena, Cilissa, and Dasypoda.) 



His genus Apis he also divides into two sections, * 

 and * * ; the first is subdivided into two families, a and 

 b (our genera Panurgus and Nomada) ; and the second 

 is divided into five subsections, a, b, c, d, e ; a and b 

 constitute families (our genera Melecta and Epeolus). 

 The subsection c is divided into two parts, 1 and 3, 

 the first containing the two divisions a and /3, each 



L 2* 



