408 



THE AMERICAN NATURALIST [Vol. LI 



out whether the species house-rat is really so variable, or 

 whether we are dealing with a new species, or whether 

 they are hybrids having a field-rat or tree-rat father, or 

 descendants, backcrosses from such hybrids. The only 

 way to get light on these questions, which are not only of 

 interest for economics, but for genetics as well, is the 

 making of hybrids. It is very probable that in reality 

 there does not exist anything which corresponds to the 

 dozens of rat species which can be found in all the mu- 

 seum catalogues. 



Zoologists and botanists often make short work of the 

 hybrid question, by. simply calling all intermediary indi- 

 viduals hybrids. In reality hybrids are very often inter- 

 mediary, especially when the parent species differ in a 

 great many genes. But very often hybrids show totally 

 new characters which would make them species in the eye 

 of several systematicians. 



We mated the small brown agouti house-rat of Java with 

 a large yellow, rather long-haired male, descending from 

 a complicated cross combining Mus rattus, Mus alexan- 

 drimis and Mus tectorum. The hybrids are dark grey, 

 with white belly and orange-ruddy sides, and very much 

 smaller even than house-rats of the same age. Rats like 

 these from a warehouse or from a ship, especially a litter 

 of similar ones as in our case, would certainly have ob- 

 tained a new species name in a museum. It is not impos- 

 sible that similar animals with a similar origin are 

 already present in a museum under a new name, as repre- 

 senting a rare species. As long as we had no proof that 

 a new alleged species of rats were not fairly constant 

 under cultivation, and produced a not too variable de- 

 scendance, we would not accept it as a good species. And 

 even so, we would require to know whether there were 

 anything in its habits of life, or in its relation to other 

 species, wliioh warranted a belief that it would not be 

 s^Y;^^|MM^ in n few generations. For the only thing which 

 (li-tiii-!ii-li( >; ;i <i)of'ies from a variety is the automatic 

 ptMiiiant'ii.-y ot" species as compared to the relative inse- 



