3IO Alexander W. Evans, 



very scarce. Since these different conditions grade into one 

 another, it seems impossible to use them as a basis for the segre- 

 gation of M. chenopoda. 



Very striking variations are to be observed in the appendages 

 of the median scales. If the series represented in Figs. 15-19 

 is examined it will hardly seem possible at first that all can have 

 been taken from a single species. The appendages shown exhibit 

 four more or less distinct types, varying in shape, in the char- 

 acter of the margin, and in the size of the component cells. In 

 the first type, shown by Figs. 15, A-H, and 16, the appendages 

 are narrowly ovate to lanceolate, tapering gradually to an 

 acute or acuminate apex; the margin is either entire or 

 provided with one or more vaguely defined and irregular teeth; 

 and the cells are large, showing no marked differences in size 

 between the median and marginal portions. In the second type, 

 shown in Figs 15, I, and 17, A, the appendages are larger 

 than in the first type and tend to be more acuminate; the mar- 

 gin is more distinctly dentate, although the teeth are still irregu- 

 lar; and the cells are much the same as in the first type, except 

 for the fact that the marginal cells in the basal portion tend to 

 be smaller. In the third type, shown in Fig. 18, A-C, the 

 appendages are broadly, ovate and apiculate to abruptly acute; 

 the margin is entire or vaguely and sparingly dentate or crenate 

 toward the base; and the cells are everywhere large, much as in 

 the first type. In the fourth type, shown in Fig. 18, D-J, the 

 appendages have about the same form as in the third type, 

 although they sometimes taper more gradually; but the margin 

 is more irregular, varying from entire to distinctly and rather 

 closely crenate, dentate, or even lobed in the basal portion; and 

 the cells are distinctly smaller, often showing a definite decrease 

 in size between the median and marginal portions. Cells con- 

 taining oil-bodies are usually absent altogether, but one or two 

 sometimes occur, as shown in Fig. 16, G, H, J. These have not 

 been observed except in the first type of appendage. 



Since the various types of appendage are more or less char- 

 acteristic of definite regions, the first type, for example, being 

 prevalent in the West Indies and the fourth in South America, 

 the writer at first thought that distinct varieties with definite 

 geographical ranges might be distinguished, using the appen- 

 dicular differences as a basis. It soon became evident, however, 



