THE CLASSIFICATION OF BIRDS 19 



Some writers have attempted to explain away the recurrence 

 of discarded characters by ascribing the fact to a mysterious ten- 

 dency, to which the name of Atavism has been given. There can 

 be no doubt that archaic characters do sometimes recur, but they 

 are probably produced by the ordinary laws of adaptation, what- 

 ever they may be, Natural Selection, or some other force. The 

 similarity between a present character and an ancient one is pro- 

 bably only accidental. The recurrence of a discarded character 

 is probably unconnected with its previous existence, the, new 

 character being independently acquired. If this theory be correct, 

 It is unscientific to speak of the reptilian affinities of so-called 

 archaic species such as Opisthocomus, Chauna, or Stndhio. They 

 do show reptilian analogies, but the characters which are pre- 

 sumably reptilian have not been inherited from reptilian ancestors, 

 but have been independently acquired, because there happened to 

 be a vacant space in the economy of Nature in which birds pos- 

 sessing those characters were able to hold their own. 



The oldest fossil birds that have yet been discovered possessed 

 teeth. Possibly in those early geological ages all birds had teeth, 

 and have since independently lost them in every group; but the 

 number of fossil birds known to us of that horizon is so very 

 small, that it is equally probable that the few toothed birds of 

 which any trace has been discovered may represent a small section 

 of the then existing avian fauna whose descendants subsequently 

 lost their teeth or became extinct. It must be admitted that 

 Archceopteryx, Hesperornis, and Ichthyornis differ so very widely 

 from each other, that they must be regarded as a somewhat dis- 

 tant development of the original avian type. It is very easy to 

 suppose that teeth were independently lost by each group of Birds 

 at various epochs, but to imagine that any Order of Birds having 

 once lost their reptilian legacy of teeth should afterwards re-acquire 

 them is difficult to reconcile with our experience of the modes of 

 evolution. 



Without, therefore, attempting to guess at the characters 

 which the original avian type possessed, we may hazard the con- 

 jecture that the Sphasniscomorphae and the Pelargomorphse became 

 aquatic in their habits, acquired webbed feet to enable them to 

 swim, and that the ancestors of the PelargomorphaB (and possibly 

 those also of the Sphaeniscomorphae) lost their fifth secondary for 

 some inscrutable reason of which we cannot make any guess. 



Be that as it may, no quincubital bird has a webbed foot. If 



