The Principles of Palaeontology. 173 



as is supposed by the f undameatal hypothesis of the transformist 

 doctrine, there is a continuity existing between all forms, including 

 those which are extinct, this continuity should appear in the dis- 

 position and structure of all the organs in the various iyy^s oi 

 one and the same series. Consequently the gradual variations of 

 the same organ in forms sufficiently near to each other, are in 

 direct relation to the filiation of the animals or plants of the 

 gr. up in question. In order to apprehend this evolution of 

 organs, we often find it necessary to recur to the principle of 

 the unity of structural plan. The celebrated theory of Geof- 

 frey Saint-Hilaire, stripped of its exaggerations, applied and 

 restricted carefully in the limits of groups which are not too 

 extensive, thus becomes a point of departure for a method rich 

 in important results. 



From this may be seen how Comparative Anatomy can furnish 

 a solid basis for the construction of phylogenic systems. It 

 aippears, even at first sight, that the problem demands only the 

 examination of an organ sufficiently characteristic and variable, 

 and that from the relations between the forms of such an 

 organ a conclusion can be arri\ ed at as to the filiation of the 

 types themselves. If this were the case nothing would be 

 easier than to convert an artificial system into a natural and 

 consequently phylogenic classification. This is what many natural- 

 ists are still doing, attributing to organs or systems of organs 

 which they have studied, a preponderating importance, sometimes 

 employing it exclusively for the establishing of genealogical 

 trees. Unfortunately the essays in this direction are often far 

 from being in accord with each other, nor can we always explain 

 these divergencies by insufficiency of material or faults of 

 interpretation. 



We must indeed appeal to those phenomena which we have 

 already cited and which have often complicated the laws of 

 organic evolution. We know that organs may undergo parallel 

 modifications in groups sometimes widely separated from each 

 other, and much more may they in series near each other yet 

 independent, and this may give rise to phenomena of convergence. 

 In this case, if we take as the basis of our estimates one of the 

 systems of organs in question, we incur the risk of confound- 



