530 MR. H. W. BATES ON THE LEPIDOPTEKA 



I'urtlier to tlio west, the species was again extremely vana1)le, but the varieties were 

 quite different from those of Ega : iiadividuals coming very the near type occurred, but 

 not one was found quite conformaljle. M. Egaeiisis was quite absent ; but, on the other 

 hand, a new variety abounded at St. Paulo, of which there was certainly no trace at 

 Ega ; this has l)een figured and described as a distinct species, viz. 31. MazcBus (Hewits. 

 Esot. Butt. Mechaititis, fig. 8). Two other remarkable varieties also occurred — one near 

 31. 3IeN02)/uliis (Hewits. I.e. 3Iechaniiis, figs. 2, 3), and the other with a structural 

 modification in the shape of the wings, described below as 31. Olwencia. 



The way in Avhich I found this species to vary, as just described, impressed me greatly, 

 and helped much, in conjunction with other facts of a similar tendency, not only to 

 di^stroy my belief in the constancy of species, but to teach how new ones may have 

 originated. The complete set of connecting forms found showed that I had here to deal 

 with one species only, disseminated over a large area, and modified in certain districts 

 under peculiar conditions there olitaining. The varieties were of such a nature as to 

 form and colours, that it was inconceivable they could have been hybrids produced by 

 the intercrossing of two or more originally distinct species. The amount of local 

 modification exhibited was not in accordance with obvious differences in the local 

 conditions; for the species was totally changed from Ega to St. Paulo, 200 miles apart 

 and very similar in soil, climate, &c., whilst very constant on the Lower Amazons, in 

 districts 600 miles apart and very different in physical conditions. Since returning 

 to England, I have learned that 31. Polymitla again varies on the eastern slopes of the 

 Andes, whilst a cluster of remarkable varieties or local forms (some of which have been 

 described as species) are found in the Andean valleys. Some of these {31. 3Iacn)ii(s, 

 3r. 3[enophilvs, &c.) are very clearly varieties of 31. Polijmnia, like the forms found at 

 Ega and St. Paulo ; l)ut others {31. 3Iothouc, 31. 3Ieiiapis) are more sharply defined, and 

 have the appeai-ance oi' tru(> species. Now I think th(> conclusion is unavoidable, that 

 these apparently distinct species are modifications, as well as the undoubted varieties are ; 

 for we have the species in all stages of modifi(!ation — simple variation, local variety 

 scarcely distinguishable from a mere variation, com2)lete local variety, and well-marked 

 rac(; or species. Tlie forms of 31. Polynuila found in Soutli Brazil confirm this view. 

 At Rio Janeiro the well-marked race or species 31. Lysimiiia alone is found; at Bahia 

 (travelling towards the hom(> of the type, 31. Poli/miiia), 31. Li/slmnia in company Avith 

 31. Nes(pa, a ibrm exactly intermediate between 31. Pol ij inula and 31. Lysbnnla; at 

 Pernaml)uco (further northward) 31. Nescea alone occurs; at Para this form is seen no 

 more, and 31. Poltjmula in its typical dress monopolizes the field. 



These fixcts seem to tc^ach that, in this and similar cases, a new species originates in a 

 local variety, formed in a certain area, where the conditions are more favom-able to it 

 tiian to the typical form, and that a large number (jf such arc simultaneously in process 

 of formation from one variable and widely distributed species. The new species cannot 

 be proved to lie estal)lished as such, unless it be fimiul in company with a sistei- Ibrm 

 wliich has ii:id a similar origin, and maintaining itself perfectly distinct from it. Cases 

 ol' two extreme varieties of a si)ecies being thus brought into contact I )y redistribution 

 or mi^-ration, and not amalgamating, will be found to l)c nununous when the sulyect is 



