ENDOTHIA PAKASITICA AND RELATED SPECIES. 



51 



search. At the northeastern limit of E. gyrosa, Clinton (15, p. 79) 

 found only a single specimen after two years' search, and the writers 

 have looked for it in all the other New England States without find- 

 ing a single specimen. The report of E. gyrosa from Massachusetts 

 by Hitchcock (42, p. 63) has already been shown to be a probable 

 error in identification. 



FACTORS INFLUENCING DISTRIBUTION. 



HOST RELATIONS. 



Just what determines the present ranges of the species can not, of 

 course, be positively decided, but some relation to certain external 

 factors may be traced. Neither species has the same distribution as 



FIG. 3. Outline map of the United States, showing the distribution of EndotJiia fluens 

 mississippiensis and E. singularis. The dots indicate collections of E. fluens mississippi- 

 ensis and crosses indicate collections of E. singularis. 



its hosts. Quercus and Fagus are both abundant farther north than 

 EndotJiia gyrosa has yet been found, while Quercus is abundant 

 north and south of the known range of E. fluens. It may be worthy 

 of note, however, that E. gyrosa extends north as far as Liquidambar 

 is found. Perhaps more significant is the relation of the range of 

 E. fluens to that of the chestnut. As will be seen from a comparison 

 of the maps (figs. 2-4) , E. fluens is not found abundantly at any point 

 outside the natural range of the chestnut. Especially interesting is 

 the fact that the southeastern limits of E. fluens and Castanea den- 

 tata are practically coincident, for in this region EndotJiia fluens 

 was found only on Castanea, never on Quercus. This suggests the 

 possibility that Castanea may be the original and favorite host of 

 EndotJiia fluens. 



