MEMOIRS OF THE NxVTIONAL ACADEMY OF SCIENCES. 45 



Upon the whole, the Ccrntocnmp'uJK are tropical, many more species oceiirrin.n' in Brazil and 

 Central America tliau in North America, aud this may be said of the family ilemileucida-. 



The family Saturniidw is a tropical group, only a single genus occurring in Europe, while in 

 North America north of Mexico there are six. In trojjical America, Africa, and southeastern 

 Asia, includiug China, the species and genera are far more numerous and form a characteristic 

 feature of the fauna.. 



Another family richly developed in the tropics of South America, Africa, and Asia is the 

 extensive family of LiisiocaDipidir, many of fclieni rivaling in size the colossal Attaci, and judging 

 from a collectiou of Central African caterpillars of this grou]) in the museum of Brown University, 

 collected on the Upper Congo, their armature of spines is tlie most formidable of any of the 

 Bombyces. And here it may be observed that the most spiny forms appear to be tropical, and 

 this tends to prove that originally nearly all our spiny caterpillars appeared in warm regions, 

 while the densely hairy forms, like Ai'ctian larvaj, predominate in cool temperate regions. 



The I'si/rhida; though so richly developed in Europe, appear on the whole to be widely 

 distributed over the tropical regions, including Australia. 



The group of CochUopodidw or slug caterpillars is richly developed in Central aud South 

 America, as well as in India, but is entirely wanting in western North America, while in Europe 

 there are only two species, this paucity or absence of species being probably due to geological 

 extinction in tlie western»portions of the Old and New Worlds. 



The small family of Mcj/alopyr/idw (Lagoidaj) is confined to the New World. One genus ( L(tf/o<() 

 occurs in the eastern United States, but the species are most numerous in the forest regions of 

 eastern South America. 



The family Liparida' appears on the whole to exist in greater force in the Tropics of America 

 and Asia than in the temperate regions to the northward. 



On the other hand, the extensive group of Arctiidw and LitJiosiidce predominate in the tem- 

 perate regions, aud its species, in rare cases — a few of Arctia — extend to the Polar Eegions, only 

 one other genus, Laria, a Liparid, sharing the regions of the Arctic Circle, a species of each genus, 

 Arctia and Laria, also being Ali)ine in Euro))e and North America. 



We will proceed to analyze the Notodontian fauna of North America. 



The animals of our American continent south of the Polar Region may roughly be divided 

 Into three grand assemblages, i. e., (1) those inhabiting the northern moist and forest-clad regions; 

 (2) those inhabiting the elevated, dry plateau region of the Cordillera mountain ranges, extending 

 southward over the Mexican plateau, aud which may be called the Plateau Province (it is Allen's 

 Arid Province) ; (3) those inhabiting the tropical portions of southern Florida and the low tropical 

 shores of southern Texas and of Central America. 



In our essay on the geographical distribution of the Geometrid moths,' published in ISlii, 

 we called attention to the elements from which our i^reseut insect fauna has been formed, and 

 claimed that the tropical elements in our fauna originally migrated fi'om Central America by 

 three avenues, i. e., the Pacific Coast, the central plateau of the Cordilleras, and the Atlantic 

 Coast, and we have always been of the* opinion that the Mexican fauna had strongly influenced 

 the Pacific Coast fauna, as well as the fauna of New Mexico, Utah, and Nevada. 



As to the Arid province, or Plateau province as it might also be designated, it may be observed 

 that within the limits of the United States it comprises the Central province of Agassiz, together 

 with the Pacific Coast or California province, and to which Dr. Allen gives the name of Campes- 

 triau subprovince. The southern eixuivalent of the Campestrian is the Mexican .subprovince. We 

 very much prefer the word Mexican to the term " Sonoran" of Dr. Merriam.^ Originally the term 

 "Sonorau" was applied by Cope to a restricted i^ortiou of northwestern Mexico known politically 

 as Sonora. 



But Dr. Merriam has, somewhat unwarrantably it seems to us, extended the term "Sonoran" 

 to include not only the elevated portions of Mexico, but also almost the whole of the United States 



'A monograph of the Geometrid moths or Phala'nida? of the United States. Report U. S. Geological Survey, 

 TT. v. Hayden, geologist in charge, Vol. X, 1876. 



-Tlie Geographic Distribution of Life in North America, with special reference to the Mammalia. Proc. 

 Biological Society of Washiugton, vii, pp. 1-64, April, 1892. With a map. 



