THE DESCENT OF THE PKIMATES 39 



tology will, of course, have a most influential 

 voice. We must hope for new discoveries to fill 

 up the immense gaps which our knowledge of the 

 Mesozoic and Palaeozoic Vertebrata yet contains. 

 And in respect to that we cannot say but that the 

 last decades have surpassed our expectations. 

 Still, it should not be forgotten that even when 

 all the fossils from those remote periods were 

 brought to light and were spread out before us, 

 they would yet remain perfectly mute with 

 respect to the details of the embryonic develop- 

 ment of the animals of which they had formed 

 part, so that on this head even posterity will 

 have to be satisfied with speculative considera- 

 tions. 



Leaving these for what they are, we may con- 

 clude by recognizing that Tarsius has taught us 

 several things : Firstly, to attach more value 

 than has hitherto been done to the inferences 

 which can be drawn from certain embryonic 

 phenomena for classificatory purposes ; secondly, 

 to entertain a certain amount of healthy scep- 

 ticism with respect to the traditional tables of 

 mammalian descent. The genera known to us 

 very rarely converge towards known predecessors 

 as we go backwards in geological time ; their 

 respective genealogies run much more parallel 



