56 PHYLOGENY 



ive example. In this pigeon we find that reduction of the checkers 

 has swept over the whole wing, leaving nothing except a few obso- 

 lete spots, which we recognize as vanishing elements of bars formerly 

 more highly developed, and homologous with those of the rock 

 pigeon. 



Here we find what at first glance looks like extraordinary variabil- 

 ity, suggesting mutations, incipient stages, bars in statu nascendi, etc. 

 The selectionist and the mutationist could each find what he looks 

 for. The first thing to decide is the direction in which the phenomena 

 are to be read. Is it a positive, progressive upbuilding of new charac- 

 ters, or a negative, retrogressive weakening of old characters? I 

 have already anticipated the answer, and will now briefly show how 

 the direction of variation is decisively settled. 



(1) These spots have every outward appearance of being reduced 

 remnants, such as we get in passing from the checkered to the 

 barred condition in rock pigeons. They are rounded or squarish 

 in form, frequently irregular and thin at the edges, dull in color as 

 if fading, etc. 



(2) The smallest stages are not found on the exposed surface of 

 the feathers, but lie concealed beneath the overlapping feathers 

 next above or in front. Concealed spots admit of but one interpre- 

 tation. This pigeon is a not distant relative of the rock pigeon, has 

 a similar gray ground, and is therefore probably moving in a parallel 

 direction, only more advanced. 



(3) The spots are found at the posterior end of the wing, near the 

 upper edge, on one to three tertials and on a few long coverts. In some 

 cases they occur also on a few of the second row of long coverts, but 

 here they are always very small and completely concealed. They are 

 thus in the position occupied by vanishing spots generally. 



(4) The adult plumage makes no advance in the number of spots, 

 and some spots (second row of long coverts) visible in the young, 

 are completely concealed in the adult. This indicates degeneration 

 unmistakably. 



(5) The stock dove, although sometimes having a concealed third 

 bar of few spots, never appears in checkered dress. It seems to have 

 moved so far in the opposite direction that no reversal of course is 

 now open to it. 



Taking the checkered pattern as the earlier one, the various con- 

 ditions of checkers and bars in rock pigeons, domestic races, and, 

 indeed, in all the wild pigeons, become almost self-explanatory. 

 We could not explain satisfactorily how just two bars could arise 

 dc novo in one species, three in another, twelve in another, and so on. 

 The repetition of dc novo origins would become ever more incredible. 

 Making phylogeny our guide as to the starting-point, we find it 

 comparatively easy to thread our way through the maze of patterns 



