THE TURTLE-DOVE PATTERN IN OTHER ORDERS OF HlllDS. 123 



lower breast, it is present in all three, but is very minute in i 2 and plainest in in 2. 

 The edge is clearly present in 3, of i, n, and of in (the black crescent absent in 

 in 3). It is present also in 4 of i and n and in in (the black crescent is absent in 

 in in 4). 



In all of these feathers from the under side of the body we see Hacker's three- 

 segmented feather. It is essentially a simple type of barring, in which we have not 

 three but four segments. 



It is difficult to decide in the second series, c.c.c., whether the "primitive" 

 yellow edge is present. It may be wholly a secondary yellow edge. The lesser 

 coverts are all marked like c. To settle this question, I have examined another 

 young (iv of pi. 56), corresponding to in (May 1903), and have selected four scapu- 

 lars from the outer edge (right side) of scapulars, beginning anteriorly (iv 1) and 

 running back to the posterior scapular (iv4). The anterior one (iv 1) has the 

 "apical light edge" plain, and about 0.5 mm. wide; the dark crescent is about 1.5 

 mm. wide; the shaft-streak cuts its way into about one-third of the dark crescent. 

 In iv 2 the white streak just cuts through the black crescent, which is reduced in 

 width, while the light (buff) edge is increased in width. In iv 3 the black crescent 

 is still narrower and the light edge still wider. In iv 4 the black crescent is still 

 narrower and the light edge still wider. 



In this last series (from right side) we see that the apical light edge is increased, 

 not at the expense of the median streak, but at the expense of the black crescent. 

 Now we may say with certainty that in the coverts shown in plate 57 (c.c.c. and 

 d.d.d.) the apical light tip does include the original light edge plus the bleached 

 black crescent. In all of these feathers the color of the apical edge is buff, the 

 streak is white, just tinged with buff. In iv of plate 56 the contrast is diminished, 

 but it is not lost, and all parts are still present. 



On examining the lesser coverts of this specimen I find that they confirm this 

 result, as they stand between the more advanced "scapular type" and the type 

 shown in c.c.c. (pi. 57). The type advances in the coverts from anterior to poste- 

 rior, the latter being most advanced (d.d.d.). It is therefore abundantly clear that 

 the feather is primarily four-segmented, two light bars and two dark bars, light 

 and dark alternating. 



The color-pattern of the robin is not an isolate pattern. Analogous conditions 

 are readily found in the common pigeons, in hybrids of the guinea-pigeon and 

 common pigeon (pi. 58), and in the American quail (Colinus virginianus) , as is 

 shown in plate 60. 



Pattern in the flicker. In Colaptes auratus we find a relatively highly modified 

 pattern, but one which has plainly departed from the turtle-dove type. The resem- 

 blance of the simpler conditions found in a series of feathers selected from an adult 

 male flicker (pi. 59), with the conditions still found in the white-winged pigeon 

 (Leucosarcia picata, pi. 44), is quite marked. The same series of feathers show that 

 in this form, too, the original pattern, in contradiction to Eimer's view, is best pre- 

 served posteriorly and most modified anteriorly. 



From the Juvenal flicker further evidence for the same conclusion is easily ob- 

 tained. The primaries, secondaries, tertials, and long coverts 19 are most instructive. 



19 A group of these feathers was arranged for the photographer, and several feathers accurately described. If a 

 photograph was prepared, it is not available to the editor. Excerpts from the explanation of this plate, describing 

 a male bird whose age is given as "one week after leaving the nest," are given above. ED. 



