THE ORIGIN AND RELATIONSHIP OF THE ROCK-PIGEONS. 51 



it at first a sudden, sport-like departure? or was it a slow and continuous transfor- 

 mation of a progressive or retrogressive nature? 



Then we come inevitably to the deeper question, which natural selection only 

 partially penetrates — the question how variation, multifarious and undirected, 

 without the aid of design or a designer, can advance to such definite and wonderful 

 achievements as specific characters. 



As the bars and chequers represent the two known extremes of a lineal series 

 of variations, if we could determine which came first in time, the problem would 

 be fairly within grasp, for then the direction of modification would be clear, and 

 the mode of progress or retrogression would be but a matter for simple observation 

 to settle. 



Such series can usually be read in either direction with about equal facility, 

 and even if all the intermediate terms are before us, the direction of sequence 

 is still ambiguous. This ambiguity infects many a biological problem, and in 

 the field of color-patterns it is omnipresent. 



In the present case it can be eliminated, or reduced to negligible limits, by 

 taking bearings from a number of different vantage-grounds. 



Which is more primitive in the rock-pigeons, the chequered or the uniform color? 

 The following detailed description of the chequers and bars of a bird representing each of 

 these two forms assist in supplying the answer. 3 



One of three chequered individuals received in 1898 from Mr. Chatwin, of Dover, was 

 less thickly chequered than the two others, which were thickly so marked. I found in the 

 least-chequered bird, which is here described, only eleven feathers, so that at least four or 

 five are missing; probably in molt. The measurements are omitted, as the series is not 

 complete. 4 One or two points, however, are worthy of notice: 



(1) The spots here are longer than in the uniform gray individual (pi. 2), and pointed 

 behind. 



(2) The gray tips are here as much shorter than in plate 2 as the spots are longer — 

 the spots being about twice as long and the tips about one-half as long. The spots being 

 at near the mid-length of the feather in both sets, the gain in length in this set, or the loss 

 in length in the other, is wholly an elongation towards, or retreat from, the tip. 



Assuming, as I believe one must, that the elongated pointed spot is the more primitive, 

 we have the means of judging whether the uniform gray with two bars of more or less squarely 

 cut spots, or the chequered variety with its second bar always more or less serrate, stands 

 nearer the original type. I do not hesitate to place the chequered variety as the earlier, 

 for its spots are of the earlier pointed form which is so generally characteristic of the 

 chequers on the anterior coverts. The uniform gray with two bars is a derived, later pattern, 

 and not, as supposed by Darwin and others, the ancestral form, from which the chequered 

 type is a secondarily derived form. This view is fully borne out by all the facts we can 

 obtain : 



(1) The most highly developed bar is the first (posterior). It is in this bar that we find 

 the greatest width and the greatest depth of color. Passing from behind forward, we find 

 the bar-spots less and less developed. This is true in all species with bars; e.g., in crested 

 pigeons the posterior bar is the most highly colored, the second bar is less brilliant, while 

 the third is still less so, and those anterior to the third grow narrower and duller. 



(2) The posterior bar seldom shows traces of a serrate edge; the second bar shows all 

 stages of serration, from the long, sharp point to the obtuse, rounded, slightly convex, and 



3 The following pages (MSS. A . r >0), in smaller type, have been inserted at this point by the editor. 



* An illustration of the chequers and bars of this bird was prepared, but it is not available to the editor. 



