THE ARMADILLO GROUP. 57 



north and south of the Mexican table-land, Seemann 1 says: "The armadillo, for in- 

 stance, which indisputably belongs to South America, is found in no part of Panama ; 

 but again appears in the neighborhood of Mazatlan, in latitude 23° 12' 0" north." 



This is probably incorrect. A recent letter from Dr. J. A. Allen, Curator of the 

 Department of Mammalogy and Ornithology, American Museum of Natural History, 

 states that two species belonging to different genera are found on the Isthmus 

 of Panama at the present time, one known as Cabassous centralis, the other being 

 the wide-ranging nine-banded armadillo, Tatu novemcinctum. The armadillo was 

 evidently an object familiar to the ancient potters of Chiriqui. As it now inhabits 

 the Isthmus, it did so, no doubt, on the occasion of Seemann's voyage in 1853. 

 One of the two species above mentioned, Tatu novemcinctum, which has " an ex- 

 tensive range from Texas to Paraguay," 2 is recognized in the ceramic art of 

 Chiriqui by the treatment of the carapace, which is composed of three quite 

 distinct regions, the banded region differing in structure from the anterior and 

 posterior sections. This triple division of the carapace is faithfully reproduced in 

 many instances. It may be done by three incised fillets (see figs. 77 and 79). A 

 more exact delineation, however, is arrived at in painted ware, when the banded 

 region is represented by transverse parallel lines and the anterior and posterior 

 division by cross lines or spots (see PI. XLI, fig. a). In one instance (see fig. 270\ 

 the middle section is composed of three bands in relief, painted red ; while over 

 the shoulders and the hips, the smooth surface of the carapace is painted black. 

 This specimen more nearly resembles the small Tatu tricinctum of South America 

 than it does the larger nine-banded species. 



That Cabassous centralis also served the ancient potter as a model is highly 

 probable. The three regions of the carapace are almost as distinct as in the nine- 

 banded armadillo, but Cabassous is not quite so large and its head, ears and tail 

 are proportionately shorter. 



In the development of art, it would be difficult to estimate how far realism lies 

 from the very first steps toward imitation on the one hand, and from the last 

 stages of conventionalism on the other. The stages at either end of the line 

 may often appear to be identical. It is therefore hazardous to decide whether a 

 given representation be a first attempt at copying a model or the work of a man 

 with a background of art inheritance and training, who can suggest the whole 

 by a skilful elimination of non-essentials. 



If the line of art development were plotted, it would probably be found to rise 

 rather suddenly to the acme of realism, and then drop slowly to about its original 

 level. The accompanying series of illustrations, however, does not begin at the 

 beginning but rather at the crest of the realistic wave, and descends gradually 

 to the trough, probably that one lying on the conventional side ; yet some of 

 the stages shown might just as well be steps in the ascending, as in the descend- 

 ing scale. In other words, a definite chronological sequence has not yet been 

 established. There are reasons, however, for the belief that a representation of 

 the entire armadillo came first, after which certain parts, such as the carapace, 



1 Op. cit., I, 261. 



2 Flower and Lydekker. An introduction to the study of mammals, 201, London, 1891. 



Memoirs Conn. Acad., Vol. III. 8 



