The pattern of pigmentation in birds is dependent on the migration and differentiation of a population of neural crest cells that develop into melanoblasts. On the basis of previous grafting experiments Rawles (1948) concluded that the pigment pattern of the chimaera is determined by the genotype of the donor melanocyte. This led Wolpert (1981) to suggest that melanoblasts from one bird can read the positional value of the ectoderm in the feather papillae of another bird. An alternative view is that an isomorphic prepattern in the feathers determines the pigment pattern. We have examined these ideas in relation to the local pigment patterns of the embryonic quail wing, distal to the elbow, where several rows of feather papillae are consistently unpigmented.

Melanin pigment is first seen at stage 35. By stage 39 a characteristic pigment pattern has been established. Most of the dorsal feather papillae are heavily pigmented, whereas many ventral papillae are unpigmented. Of the ventral papillae three rows (E2, E3 and H2) are always unpigmented, and it is these three rows that form the basis of the quail local pattern. The DOPA reaction indicates that no melanoblasts are present in these white feathers, although they are present in all the feathers of the White Leghorn wing.

When quail neural crest cells are grafted to the chick, either isotopically or to the wing bud, all or nearly all rows of ventral papillae become pigmented by stage 39. The only evidence of donor influences in the pattern is that, in some grafts, rows E2–3 have a high proportion of unpigmented papillae, and wings from earlier stages resemble the quail. When unpigmented papillae are present, histology shows that they contain undifferentiated crest cells. When introduced into a quail wing bud, chick crest cells enter all the feather papillae of the wing, including those in rows E2–3 and H2.

We suggest that neither the positional information nor the prepattern theory alone can account for all of our findings. Contrary to previous claims, local cues may be important in determining crest-cell differentiation. We have established that crest cells migrate into all feather papillae of the quail-chick chimaera, including those that will remain unpigmented. We show that neither differential migration nor differential proliferation is involved in pattern formation in the quail-chick chimaera.

The pattern of pigmentation in birds is remarkably varied and beautiful. It is not only of great intrinsic interest, but provides a system for investigating pattern formation and its relation to positional information (Wolpert, 1969); it involves the migration, multiplication and differentiation of just one cell type, the neural crest cell, in a varying environment.

The blacks, greys, browns and rufous colours of bird feathers, as well as the more subdued shades of yellow, are produced by melanin pigments synthesized by epidermal melanocytes (Rawles, 1960). These cells originate in the neural crest (Dorris, 1939), and follow the dorsolateral migratory pathway before entering the ectoderm (Teillet, 1971; reviewed by Le Douarin, 1982). Two theories can be put forward to account for pigment-pattern formation, and they have very different predictions for the pattern which one would expect to see in a neural crest chimaera.

Rawles (1948) grafted pigment cells between bird embryos; she concluded that ‘irrespective of the immediate source of the melanoblasts or the method of introducing them into foreign feather germs, the results have been consistent in showing that the melanophores produce their specific color and pattern in homologous feathers of varieties normally exhibiting an entirely different color and pattern’ (Rawles, 1948). This led Wolpert (1981) to suggest that melanoblasts read the positional value of the feather ectoderm and interpret this according to their own genome (Fig. 1A). The key point is that the positional values in host and donor are identical, and the pattern arises from the differential response of the melanoblast. The prediction from this theory is that when pigment cells are grafted between species, the pattern in the chimaera will resemble the donor, if melanoblasts can read the local positional values.

Fig. 1.

Diagram illustrating two models to account for pigment-pattern formation. (A) The positional information model. (B) The prepattern model. Black circles=pigmented feather papillae; open circles=unpigmented papillae.

Fig. 1.

Diagram illustrating two models to account for pigment-pattern formation. (A) The positional information model. (B) The prepattern model. Black circles=pigmented feather papillae; open circles=unpigmented papillae.

A recent study of the quail-chick chimaera has provided further evidence in favour of this theory (Kinutani and Le Douarin, 1985) even though the chimaera is interspecific. Yet more evidence comes from grafting experiments with amphibia. Donor-type patterns are seen in amphibian chimaerae provided the species are not too distantly related (DuShane, 1943; Rawles, 1948). Twitty (1936, 1937) and Twitty and Bodenstein (1939) grafted neural crest cells between different species of Triturus, and between Triturus and Ambystoma. They found that the pattern in the chimaera was determined largely by the genotype of the grafted neural crest.

An alternative theory is based upon the pattern being generated by an isomorphic pre-pattern (Wolpert and Stein, 1984) in the feather papillae (Fig. 1B). In this model there are local differences that determine whether or not a feather becomes pigmented. Thus, in this case, the pattern is determined largely by the special properties of the host feather germs to which the melanoblasts can respond. In principle, melanoblasts from any species capable of this response would produce a host-type pattern when grafted to another bird.

Neither theory says anything about the cellular or molecular mechanisms involved; indeed, little is known about these. It could be that differential migration of melanoblasts is involved. Thus Watterson (1942) and Rawles (1959) have suggested that white feathers arise because of some block to crest-cell entry. It is important to remember that these workers had no reliable histological markers for identifying crest cells. Alternatively, pigment cell precursors could migrate to all parts of the skin, but only differentiate in certain areas. Although there is little evidence from birds to support either view, differential migration does seem to be important in pigment patterning in embryonic and larval forms of fish (Trinkaus, 1988) and amphibia (Epperlein and Claviez, 1982).

The White Leghorn has long been used as a host both in pigmentation studies (Rawles, 1948) and in studies of the quail-chick chimaera (Le Douarin, 1982). Jimbow et al. (1974) have made a detailed study of pigmentation in this breed. Crest cells enter the feather papillae and give rise to large numbers of DOPA-positive cells (melanoblasts). These cells are present in the feather germs mainly during the period of 9– 16 days of incubation. These melanoblasts may differentiate into melanin-containing cells (melanocytes), although they die before depositing significant amounts of melanin in the feather papilla. These short-lived melanocytes can be seen in the basal part of feather papillae where they form a black ring (Rawles, 1944). The defect that causes premature death in the melanocytes is intrinsic to the melanocytes themselves (Hamilton, 1940) and only affects those of the neural-crest lineage (thus the retina is pigmented in this breed because retinal melanocytes arise from the optic cup and not the neural crest). On the basis of genetic studies Hadley (1915) found the White Leghorn to possess latent pigment-patterning genes, and described this breed as ‘a masquerader who conceals many colors and patterns beneath her pure white plumage’. Brumbaugh (1967) considers the White Leghorn to have the Extended black (E) pigmentation genotype, a genotype it shares with such breeds as the Black Australorp, Black Minorca and Barred Plymouth Rock.

While the experiments of Rawles (1948) support the positional information model, several crucial points remain unresolved. It has not been shown that local patterns in the donor are seen also in the chimaera. By local pattern we mean a characteristic arrangement of pigmented and unpigmented feather papillae in a defined area. The ‘donor patterns’ described in the literature to date refer merely to the spots and stripes within single feathers. Our study is of grafts between the Japanese quail and the White Leghorn chick, and there are two reasons for this choice of breeds. First, there is no detailed description in the literature of the development of pigment patterns in the quail-chick chimaera. Second, most grafting experiments to study pigmentation in birds have been of the chick-chick type; it is not at all clear what happens in xenografts. Rawles (1939) found that robin pigment cells grafted to the White Leghorn did not produce the appropriate donor pattern, whereas Willier (1941) found that xenografts using a variety of bird species all resulted in donor coloration. Once again, however, difficulty arises because these authors appeared to be looking simply at patterns and colours within feathers, and not at local patterns. Here we present a study of pigment-pattern formation in the quail embryo, in the White Leghorn chick, and in heterospecific combinations between these two breeds.

Fertilised eggs of the Japanese Quail (Coturnix coturnix japonica) were obtained from a colony kept in this department. Fertilised eggs from a strain of the White Leghorn (‘Ross White’) were obtained commercially (Needle Farms, Hertfordshire).

Mapping the distribution of pigmented and unpigmented feather papillae on the wings of the normal quail and the operated chick

Feather papillae were plucked from the wing of the right-hand side, and the position of each papilla was represented on a chart with a note of its pigmentation type. Only distinct elevations greater than 0.2mm in height or diameter were recorded. Simple ectodermal thickenings were ignored. Three categories of pigmentation were designated according to the appearance of the papilla under the dissecting microscope: ‘pigmented’, ‘unpigmented’ and ‘trace’. A papilla is categorised as ‘trace’ when its melanin is visible only at high power (30x) under the dissecting microscope. The distinction between these categories is subjective, but they allow a large number of feather papillae to be graded in a relatively short space of time.

Grafts of quail neural-crest cells to the chick embryo

Two methods were used

  1. Isotopic grafts of quail neural hemi-tubes with associated neural crest, 2 somites in length, were made to chick embryos. Our procedure is a modification of that illustrated by Le Douarin (1982, p. 14). Donors and hosts were incubated for 40–48 h at 38°C±1°C, and had between 11 and 22 somites at the time of grafting. Tissue for grafting was excised with tungsten needles; trypsinisation was not used. The hemi-tubes were produced by cutting the neural tube longitudinally into left and right halves. The corresponding piece of neural tube in the host was not divided longitudinally, but was removed whole. The grafts were made in most cases at the level of somites 15–20.

  2. Grafts of whole, quail neural tubes and associated neural crest, 2 somites in length, were taken from the thoracic region of a quail embryo and inserted into the wing bud of a chick embryo. The procedure is essentially that of Dorris (1939). The donors were incubated for two days at 38°C±1°C and had from 14 to 18 somites at the time of grafting. The hosts were of stages 16–21 (Hamburger and Hamilton, 1951). A slit, parallel to the long axis of the embryo, was cut in the base of the right wing bud with tungsten needles and the graft worked into the slit so that no tissue was left protruding. With younger hosts, the graft sometimes passed through into the coelom. All embryos were incubated for a further 9–17 days before killing.

Grafting quail wing buds to the chick

The right wing bud was removed from a chick, and a quail wing bud grafted in its place. The grafted wing bud was held in place by platinum wire pins. Donors and hosts were of the same age within the range of stages 17–24.

The DOPA reaction

The tissue was fixed for 2h in 10% formalin in phosphate-buffered saline (PBS, 4°C, pH7·4) then rinsed in PBS for 1 h. The reaction mixture consisted of 0·1 % D-L dihydroxyphenylalanine (DOPA; Sigma) in PBS (pH7·7) which had in some cases been gassed overnight with nitrogen. Incubation was carried out at 37°C for 6–14 h during which time the reaction mixture was changed twice. Controls were treated as above, but DOPA was omitted from the reaction mixture and in some cases 10−3 M-sodium diethyldithiocarbamate (Sigma), a tyrosinase inhibitor, was included.

Histology

Tissue was fixed overnight in half-strength Kamovsky’s fixative (Kamovsky, 1965), dehydrated in graded alcohols and embedded in Araldite. Sections were cut at 2μm and stained with toluidine blue or the Feulgen-Rossenbeck technique.

The arrangement of feather papillae in the quail and chick

The maps presented here are diagrammatic rather than topographic representations of the wing surface. The feathers of the wing can be readily resolved into a series of rows roughly parallel to the proximodistal axis of the wing. These rows are arranged as discrete groups (tracts) as can be seen in Fig. 2. The basic unit of our maps is not the homologous papilla but the homologous row. It was not found to be possible to produce a standard template for comparing homologous papillae on different wings. The reason for this is the variability in the pterylosis of different wings; not only does the number of papillae per row vary from wing to wing at a particular stage, but the number of rows in certain tracts varies also.

Fig. 2.

Quail wing (stage 37). (A and B) Ventral surface; (C and D) dorsal surface. Scale bar on A=2mm.

Fig. 2.

Quail wing (stage 37). (A and B) Ventral surface; (C and D) dorsal surface. Scale bar on A=2mm.

Our maps include most of the feather papillae distal to the elbow except those at the wrist, and any other papillae that are difficult to assign accurately to rows. We have given a letter and number to each row of papillae: the letter denotes the group (tract) to which that row belongs, and the number indicates the position of the row within the tract (Fig. 2). The most posterior row in each tract is number 1, the next row number 2, and so on.

The tracts containing the B and G rows are difficult to map accurately. These tracts are both essentially triangular and contain more rows at their broad (proximal) end than at their pointed (distal) end. Additionally, B and G rows often bifurcate or end suddenly in mid-course. The B tract is particularly problematic since it merges at an angle into the A rows, and the boundary is not always easy to establish.

Some rows are highly consistent: rows El–4 and Hl–2 show interesting local pigment patterns, and are amongst the best-defined rows of the wing. For these reasons, the E and H rows will form the basis of our report. Our row designations compare with the nomenclature of Lucas and Stettenheim (1972) as follows: El and Hl are under major coverts; E2 consists of under median coverts; and E3–4 and H2 are under minor coverts; the ‘downs’ described by these authors are not included in our account, and must not be confused with other feathers near the posterior wing margin. Rows El and Hl are continuous with each other; so too are rows E3 and H2. Rows E2 and 4 have no counterparts in the hand. Row Fl is the most posterior row lying between the E and G groups. There are other papillae in this region but they are not included in this account because they do not always form recognisable rows. The I group usually consists of 3 rows. If a fourth is present this is disregarded so that 13 is always the long row on the anterior margin of the ventral hand.

Fig. 3 shows light micrographs of typical papillae from each category. There is clearly some overlap between categories, although we have found that most papillae fall clearly into either the ‘pigmented’ or ‘unpigmented’ class.

Fig. 3.

Light micrographs of wholemounted feather papillae illustrating the three different categories of pigmentation. (All are from the wing in Fig. 10.) (A) Papilla scored as ‘unpigmented’; (B) scored as ‘trace’; (Cand D) both scored as ‘pigmented’. All are to the same scale as A. Scale bar in A=200μm.

Fig. 3.

Light micrographs of wholemounted feather papillae illustrating the three different categories of pigmentation. (All are from the wing in Fig. 10.) (A) Papilla scored as ‘unpigmented’; (B) scored as ‘trace’; (Cand D) both scored as ‘pigmented’. All are to the same scale as A. Scale bar in A=200μm.

The normal pigmentation of the quail wing

We have applied chick stages to the quail; beyond stage 39, however, this cannot be done accurately. For this reason, stages 39–41 have been grouped together as a single category.

Stages 35–38

Traces of melanin may be seen in the incipient papillae around the elbow as early as stage 35. However, melanocytes are not seen in significant numbers until stage 36 (4 cases) on the dorsal surface of the wing at the proximal ends of rows Al–4. During stages 37 (3 cases) and 38 (3 cases), pigment appears in successively more anterior and distal papillae on the dorsal surface (Fig. 2). On the ventral surface, papillae appear at stage 37, and pigment at stage 38, at which time it appears in the newly formed rows Gl-2, El and E4.

Stages 39–41 (6 case

By this time, the definitive pattern has been established. Fig. 4 shows a quail wing of about stage 40. All the dorsal papillae are heavily pigmented with the exception of rowB5, and the distal members of DI. The ventral surface has developed a characteristic pattern: rows E2–3, G5 and H2 are consistently unpigmented, whilst the remaining rows show varying amounts of pigment. Since, as mentioned above, the definition of the G rows is problematic, it is the H and I rows that contain the most important local patterns. In rows Gl–3 and 12–3 the density of pigmentation approaches that seen in dorsal feather papillae; this gives the appearance on the ventral surface of a streak of pigmented papillae running proximodistally along the anterior margin of the wing (Fig. 4A). Stages 38 and beyond show a graded decrease in the frequency of pigmentation across the B rows. The data are shown in Figs 56, where histograms show the frequency of pigmented papillae in each row. We have examined, but not mapped, wings from quails up to 10 months after hatching; the wing plumage of these birds shows broadly the same pattern as that shown by stage 39–41 embryos.

Fig. 4.

Quail wing, about stage 40. (A and B) Ventral surface; (C and D) dorsal surface. Scale bar on A=5mm.

Fig. 4.

Quail wing, about stage 40. (A and B) Ventral surface; (C and D) dorsal surface. Scale bar on A=5mm.

Fig. 5.

Histograms illustrating the development of pigment in the papillae of the ventral surface of the quail wing. The abscissa shows the row number, and the ordinate shows the mean percentage, from all the cases of that stage, of pigmented papillae in each row. Error bars show the standard error of the mean. (A) Stage 38; (B) stage 39–41.

Fig. 5.

Histograms illustrating the development of pigment in the papillae of the ventral surface of the quail wing. The abscissa shows the row number, and the ordinate shows the mean percentage, from all the cases of that stage, of pigmented papillae in each row. Error bars show the standard error of the mean. (A) Stage 38; (B) stage 39–41.

Fig. 6.

Histograms illustrating the development of pigment on the dorsal surface of the quail wing. (A) Stage 36; (B) stage 37; (C) stage 38; (D) stage 39–41.

Fig. 6.

Histograms illustrating the development of pigment on the dorsal surface of the quail wing. (A) Stage 36; (B) stage 37; (C) stage 38; (D) stage 39–41.

Since we are to concentrate on the E and H rows, we have examined these in a further 11 embryos, making a total of 17 cases of stages 39-41. None showed any pigment in E2-3 or H2. Feathers in El and Hl are black or grey at the base only, or base and tip only, with white intervening. One third of cases show no pigment at all in E4. The intensity of pigment in a row generally declines towards its distal end. Black only, and not brown, is seen in the E and H rows. The only exceptions are the occasional embryos whose entire plumage lacks black melanin; in these cases, brown replaces black in all feathers.

14 quail wings,, from stages 36–41, were examined with the DOPA reaction. The other wing from each embryo was untreated. A positive reaction appeared to be confined to those papillae that were already in the process of becoming pigmented; thus there was little difference between treated and untreated wings. In no cases were any DOPA-positive cells found in the unpigmented rows (E2–3, H2) of the ventral surface.

Pigmentation of the White Leghorn

The plumage appears entirely white to the naked eye, although punctate or fusiform melanocytes can be seen under the microscope in cleared specimens. We have examined 15 wings from stages 37–45. Melanocytes are seen first in the larger papillae on the dorsal surface, and in the skin of the cubital apterium, at the end of stage 37. Of the 9 wings examined of stages 37–39, 6 had no melanocytes at all, but by stage 42, all wings showed melanocytes at the bases of all their papillae. Fig. 7A shows a papilla (row E3) from a normal stage 39 White Leghorn. Immature punctate melanocytes are seen in the skin, in a ring at the base of the papilla, and scattered up the shaft of the papilla. All the feathers of the wing appear to be capable of supporting the differentiation of these cells. By stage 42 (Fig. 7B), the rings have become particularly prominent.

Fig 7.

Cleared whole mounts of normal White Leghorn feather papillae. (A) Papilla from row E3 (stage 39). Melanocytes are seen in the skin, the shaft of the papilla, and in a ring at the base of the papilla. Scale bar=250μm.

(B) Skin from stage 42 wing showing the prominent melanocyte rings (arrow). Scale bar=250μm.

Fig 7.

Cleared whole mounts of normal White Leghorn feather papillae. (A) Papilla from row E3 (stage 39). Melanocytes are seen in the skin, the shaft of the papilla, and in a ring at the base of the papilla. Scale bar=250μm.

(B) Skin from stage 42 wing showing the prominent melanocyte rings (arrow). Scale bar=250μm.

The DOPA reaction reveals large numbers of mel-anoblasts in White Leghorn feather papillae (cf. Jim-bow et al. 1974). Moreover, they are present in all the feather papillae of the wing, including El–4 (Fig. 8) and Hl–2. These DOPA-positive cells are not seen if diethyldithiocarbamate, an inhibitor of tyrosinase, is present in the reaction medium. This indicates that our technique is specific for melanoblasts.

Fig. 8.

Cleared whole mount of White Leghorn skin (11 days incubation) after treatment with DOPA. Scale bar=400μm.

Fig. 8.

Cleared whole mount of White Leghorn skin (11 days incubation) after treatment with DOPA. Scale bar=400μm.

Pigmentation patterns of the wing of chicks after quail neural tube grafts

Isotopic grafts

Of 91 grafts performed, 45 survived and 27 of these were either entirely unpigmented, or showed irregular patches of pigment, usually in the feathers of the dorsal forearm. In these wings, the ventral forearm and at least the distal half of the hand are unpigmented. These cases are not included in our account since we consider that they represent a failure of the grafted quail cells to become properly incorporated into the host. This leaves 18 cases. The right wing was chosen for study wherever possible but, in some cases, only the left wing had become pigmented and was taken instead. In either case, only one wing was taken from a single embryo. In all of the cases examined, pigment spread beyond the axial level of the graft, indicating that the number of grafted cells was sufficient to populate the entire wing region.

Stages 36 (3 cases) and 37 (6 cases) resemble the quail although pigmentation appears to develop more rapidly than in quails of the same stage. A stage 37 graft is shown in Fig. 9. At stage 38 (3 cases), there appears to be a precocious expression of pigment, compared to the normal quail, on the ventral wing surface. During stages 38–41, quite a different pattern emerges whereby most of the feathers of the wing become pigmented. In wings of stages 39–41, (6 cases) virtually all of the dorsal papillae are pigmented and most of those on the ventral surface are too (Fig. 10). Nearly all the rows of the wing contain a high proportion of pigmented papillae: at least 80% in most cases (Figs 11, 12). Rows E2–3 are exceptional in that fewer than half of their papillae are pigmented (the mean percentage of pigmented papillae in these rows is: E2=36%, E3=43%). These data are the means from several cases. Row H2 is consistently unpigmented in the quail, but in the operated chick wings it is always pigmented.

Fig. 9.

Chick wing (stage 37) after isotopic graft of quail neural crest. (A and B) Ventral; (C and D) dorsal. Scale bar on A =4 mm.

Fig. 9.

Chick wing (stage 37) after isotopic graft of quail neural crest. (A and B) Ventral; (C and D) dorsal. Scale bar on A =4 mm.

Fig. 10.

Chick wing (stage 40) after isotopic graft of quail neural crest. (A and B) Ventral; (C and D) dorsal. Scale bar on A=7 mm.

Fig. 10.

Chick wing (stage 40) after isotopic graft of quail neural crest. (A and B) Ventral; (C and D) dorsal. Scale bar on A=7 mm.

Fig. 11.

Histograms illustrating the development of pigment in the isotopic graft, ventral surface. (A) Stage 37; (B) stage 38; (C) stage 39–41.

Fig. 11.

Histograms illustrating the development of pigment in the isotopic graft, ventral surface. (A) Stage 37; (B) stage 38; (C) stage 39–41.

Fig. 12.

Histograms illustrating the development of pigment in the isotopic graft, dorsal surface. (A) Stage 36;(B) stage 37; (C) stage 38;(D) stage 39–41.

Fig. 12.

Histograms illustrating the development of pigment in the isotopic graft, dorsal surface. (A) Stage 36;(B) stage 37; (C) stage 38;(D) stage 39–41.

To give some idea of the variability between grafts, the E and H rows were examined in a further 5 grafts making a total of 11 grafts of this stage. Rows E2–3 were entirely pigmented in 4 cases, entirely unpigmented in 2 cases, and showed a mixture of pigmented and unpigmented papillae in 5 cases. Row H2 always contained pigmented papillae.

For histological analysis, 44 grafts were performed providing 29 cases. The early phase of crest-cell migration in the chimaera, including the seeding of the ectoderm, has been considered by Teillet (1971). We consider stage 33 onwards. Transverse sections of the forearm show quail cells in the ectoderm, the loose connective tissue beneath the dermis and in neurovascular bundles. From stage 37 onwards, a network of pigmented cells is formed which extends round the full circumference of the limb under the dermis, and between the blocks of skeletal muscle of the limb. Quail cells were at all stages rare in the dermis, and were never seen in cartilage or in the dermal pulp of feather papillae. Quail cells became distributed to the ectoderm round the entire circumference of the limb. We could find no evidence to support the idea that crest cells migrate more readily into pigmented papillae than into unpigmented papillae, or that they proliferate more in pigmented papillae.

In 4 wings of stage 39, the proportion of ectodermal cells bearing the quail nucleolar marker was determined. In the heavily pigmented basal part of El papillae, this proportion was in the range 1/1243—1/296, and all of the quail cells were pigmented. Melanocytes were always located in their definitive position at the apex of the barb ridges. The characteristic quail nucleolus becomes dispersed and faint as the melanocyte matures. Therefore, to see if we were underestimating the number of quail cells, we did some more counts scoring not for the nucleolar marker, but simply for melanocytes. Our criteria were that the nucleus be completely surrounded by melanin granules, but not overlaid by them. This produced values of the same order of magnitude, the range being 1/442–1/302.

Next, papillae from rows E2–3 were examined. They were either entirely unpigmented, or had only slight traces of pigment at their bases. Quail cells, always unpigmented, were seen in all the papillae and the frequency was in the range 1/4068–1/432. The mean (1/666) is in the range of the frequency of quail cells seen in pigmented papillae. The unpigmented quail cells were always located in the definitive melanocyte position at the apex of the barb ridges (Fig. 13).

Fig. 13.

Light micrograph showing a transverse section through an unpigmented papilla (row E3, isotopic graft, stage 39). (A) Phase contrast. Scale bar=20μm. (B) Inset shown in 13A, bright-field. Note the unpigmented quail cell at the apex of the barb ridge (arrow). Scale bar=10μm.

Fig. 13.

Light micrograph showing a transverse section through an unpigmented papilla (row E3, isotopic graft, stage 39). (A) Phase contrast. Scale bar=20μm. (B) Inset shown in 13A, bright-field. Note the unpigmented quail cell at the apex of the barb ridge (arrow). Scale bar=10μm.

Grafts of quail neural tube to the chick wing bud

22 grafts were performed, all of which survived, providing 8 cases of stages 39–41 for examination. The remainder were only partially pigmented and are not included in our analysis. The wing shown in Fig. 14 has a similar pattern to that of the isotopic graft at the same stage (cf. Fig. 10). A comparison between Figs 11C and 15A shows that in both cases all ventral rows, includingrows E2–3 and H2, contain pigmented papillae (E2=30%, E3=23% and H2=86%). As was the case with isotopic grafts, the mean frequency of pigmented papillae in rows E2–3 is lower than in all of the other rows. Thus a similar pattern is produced by quite a different grafting technique.

Fig. 14.

Chick wing (stage 39–40) after a graft of quail neural crest into the wing bud. (A and B) Ventral; (C and D) dorsal.Scale bar on A=7 mm.

Fig. 14.

Chick wing (stage 39–40) after a graft of quail neural crest into the wing bud. (A and B) Ventral; (C and D) dorsal.Scale bar on A=7 mm.

Grafts of quail wing buds to the White Leghorn chick

28 grafts were performed, 5 of which died leaving 23 cases. In 16 cases, the donor quail was of stage 16–20, and 15 of these developed into completely unpigmented wings. One case (donor stage 18) had a few greyish papillae on the dorsal forearm but was otherwise unpigmented. Of 5 cases that used stage 21 donors, 2 developed into unpigmented wings, and 3 into pigmented wings. A stage 22 wing bud, and stage 25 wing tip, both developed into pigmented wings. Thus it seems that neural crest cells first enter the wing of the normal quail at stage 21 or later. Unpigmented quail wings from this series were treated with the DOPA reaction. Of 12 cases treated (stages 38–39), 4 showed melanoblasts in all the feather germs of the wing, including rows El–4 and Hl–2 (Fig. 16). Histology, coupled with Feulgen-Rossenbeck staining, confirmed the view that these cells were chick melanoblasts that had migrated into the quail wing. Five wings showed melanoblasts distributed in irregular patches, as though the chick cells had failed to enter the wing in sufficient numbers. A further three wings of this type had diethyldithiocarbamate included in the DOPA reaction mixture. These showed no staining at all, which suggests that the reaction is specific for melanoblasts.

Fig. 15.

Histograms illustrating the development of pigment in the chick wing following a graft of quail crest to the chick wing bud. (A and B) both=stage 39–41.

Fig. 15.

Histograms illustrating the development of pigment in the chick wing following a graft of quail crest to the chick wing bud. (A and B) both=stage 39–41.

Fig. 16.

Unpigmented quail wings produced by grafting quail wing buds before stage 21 to the White Leghorn. (A) Treated with DOPA and diethyldithiocarbamate (gassed with N2), stage 38. Scale bar=2mm. (B) Treated with DOPA only (stage 38). Scale bar=2mm. (C) Cleared wholemount from B (boxed area) showing chick melanoblasts in the papillae of the E group.

Fig. 16.

Unpigmented quail wings produced by grafting quail wing buds before stage 21 to the White Leghorn. (A) Treated with DOPA and diethyldithiocarbamate (gassed with N2), stage 38. Scale bar=2mm. (B) Treated with DOPA only (stage 38). Scale bar=2mm. (C) Cleared wholemount from B (boxed area) showing chick melanoblasts in the papillae of the E group.

The normal quail embryo shows a characteristic pigment pattern in its wing plumage. The most striking features of the pattern are: (1) the consistently heavy pigmentation of the dorsal surface compared to the ventral; (2) the graded decrease in pigment across rows B1–B5; (3) the dark streak of feathers near the anterior margin of the ventral surface, formed by the heavy pigmentation of rows Gl–3 and 12–3; and (4) the consistent lack of pigment in rows E2–3 and H2. Features 1–3 are dependent on the relative amount of melanin in individual papillae; feature 4, however, the predictable absence of pigment in a given set of feather papillae, defines a local pattern.

We find that when quail neural crest is grafted isotopically to the chick, the pigment patterns produced in the chick feathers at stages 36-37 resemble those of the normal quail, although pigmentation develops somewhat more rapidly than in the quail. But by stages 39–41, the pattern in the host has become very different from the pattern of the normal quail wing. Although the dorsal surface is heavily pigmented like the quail, the ventral surface, instead of showing well-defined local patterns, shows pigment in most or all of its rows. In other words, the chick host shows a generalised increase in the frequency of pigmented papillae compared with the normal quail. Grafting quail neural crest into the chick wing bud produces in the chick plumage, at stages 39–41, the same pigmentation as is produced by an isotopic graft; that is, melanin is found in most or all of the rows of papillae. This shows that the method of grafting does not influence the result.

Although the pigment pattern of the chimaera is very different from that of the donor, there is some indication of donor pattern in two rows of papillae, E2 and E3. It will be remembered that in the quail, these two rows are always unpigmented. After both types of graft, most of the rows on the host wing at stages 39–41 show pigmentation in more than 80% of their papillae. The exceptions are rows E2 and E3 in which, on average, fewer than half of the papillae are pigmented. Indeed, in some cases, pigment is entirely absent from these rows. If these unpigmented papillae are examined histologically, undifferentiated quail cells are found to be present. Another row that is always unpigmented in the quail is row H2; in operated chicks however, this row is always pigmented. It is possible that quail crest cells can read the positional value of rows E2–3 in the chick. Kinutani and Le Dourain (1985) have reported that the dorsal stripes on the quail embryo are seen also in the chimaera, and have interpreted this as indicating that quail cells can read the positional values of chick feathers.

Our studies of the White Leghorn have shown that all the feather germs of the wing are capable of supporting at least the initial stages in the differentiation of melanocytes. Our evidence for this is of two types. First, rings of melanocytes are found at the bases of all the feather papillae of the White Leghorn wing. Second, DOPA-positive melanocytes are present in all the feather germs of the wing. In the normal quail, by contrast, only those feathers that become pigmented contain DOPA-positive melanoblasts; the unpigmented papillae on the ventral surface do not. By grafting quail wing buds of different stages to the chick, we find that crest cells do not enter the quail wing bud in significant numbers until stage 21 and later.

Our results do not support the hypothesis of Rawles (1959) and Watterson (1942) which says that white feathers arise because of some physical block to mel-anoblast entry. Our reasoning is based on two lines of evidence. Our histological analysis of the quail-chick chimaera shows that undifferentiated neural crest cells are present in white feather papillae, and their numbers in these papillae are comparable to the number of quail cells in pigmented papillae. Furthermore, we find that when quail wing buds are grafted to the chick, DOPA staining reveals that chick melanoblasts have migrated into all the quail feathers, including those in rows E2–3 and H2 which are, in the normal quail, unpigmented.

Our results cannot be explained by either the positional information or prepattem theories alone. The low frequency of pigmented papillae in E2-3 in the chimaera suggests that quail cells can read the positional values of the ectoderm in those papillae. However, this does not apply to the other papillae of the chick wing; most host papillae become pigmented, including those in H2 that are always unpigmented in the quail. In these, the quail cells may be responding to local cues in the papillae; we have shown that all the papillae of the chick wing are capable of supporting at least the early stages of chick melanocyte differentiation. When chick crest cells are introduced into the quail wing, they enter all rows of papillae and differentiate into melanoblasts. This means that either there is no prepattem in the quail wing or, if there is one, chick melanoblasts do not respond to it.

We propose the following model to account for pigment-pattern formation in the plumage of the avian embryo wing. Neural crest cells enter the wing bud at stage 21 where they subsequently become distributed to all regions of the ectoderm. Crest cells enter all feather papillae, including those destined to form white feathers, and they come to occupy the apical region of the barb ridges. Local cues in the feather papillae control crest cell differentiation such that, in pigmented feathers, dendritic melanocytes are formed whereas, in white feather papillae, the crest cells remain undifferentiated with no tyrosinase activity. Whether these local cues represent a response of the crest cell to positional information, or to special local properties constituting a prepattem remains to be determined.

The White Leghorn has been the breed of choice as a host in pigmentation studies for many years. It could, however, be argued that the lack of pigmentation in this breed leads to anomalous interactions. We are therefore investigating pigmentation patterns in chimaerae constructed from other bird species. Preliminary findings suggest that a prepattem specified by positional information can best account for most of the results.

We are grateful to Dr Cheryll Tickle for her comments on the manuscript. This work is supported by a grant from the Science and Engineering Research Council.

Brumbaugh
,
J. A.
(
1967
).
Differentiation of black-red melanin in the fowl: interaction of pattern genes and feather follicle milieu
.
J. exp. Zool
.
166
,
11
24
.
Dorris
,
F.
(
1939
).
The production of pigment by chick neural crest in grafts to the 3-day limb bud
.
J. exp. Zool
.
80
,
315
345
.
Dushane
,
G. P.
(
1943
).
The embryology of vertebrate pigment cells. Part 1. amphibia
.
Q. Rev. Biol
.
18
,
109
127
.
Epperlein
,
H. H.
and
Claviez
,
M.
(
1982
).
Formation of pigment cell patterns in Triturus alpestris embryos
.
Devi Biol
.
91
,
497
502
.
Hadley
,
P. B.
(
1915
).
The White Leghorn - a masquerader who conceals many colors and patterns under her pure white plumage
.
J. Hered
.
6
,
147
151
.
Hamburger
,
V.
and
Hamilton
,
H. L.
(
1951
).
A series of normal stages in the development of the chick embryo
.
J. Morph
.
88
,
49
92
.
Hamilton
,
H. L.
(
1940
).
A study of the physiological properties of melanophores with special reference to their role in feather coloration
.
Anat. Rec
.
78
,
525
547
.
Jimbow
,
K.
,
Szabo
,
G.
and
Fitzpatrick
,
T. B.
(
1974
).
Ultrastructural investigation of autophagocytosis of melanosomes and programmed death of melanocytes in White Leghorn feathers: a study of morphogenetic events leading to hypomelanosis
.
Devi Biol
.
36
,
8
23
.
Karnovsky
,
M. J.
(
1965
).
A formaldehyde glutaraldehyde fixative of high osmolarity for use in electron microscopy
.
J. Cell Biol
.
27
,
137a
. (Abstr.)
Kinutani
,
M.
and
Le Douarin
,
N. M.
(
1985
).
Avian spinal cord chimeras. 1. Hatching ability and posthatching survival in homo- and heterospecific chimeras
.
Devi Biol
.
111
,
243
255
.
Le Douarin
,
N. M.
(
1982
).
The Neural Crest
.
London/New York
:
Cambridge Univ. Press
.
Lucas
,
A. M.
and
Stettenheim
,
P. R.
(
1972
).
In Avian Anatomy: Integument, Part 1
.
Washington D.C
.:
United States Department of Agriculture
.
Rawles
,
M. E.
(
1939
).
The production of robin pigment in white leghorn feathers by grafts of embryonic robin tissue
.
J. Genet
.
38
,
517
531
.
Rawles
,
M. E.
(
1944
).
The migration of melanoblasts after hatching into pigment-free skin grafts of the common fowl
.
Physiol, tool
.
17
,
167
183
.
Rawles
,
M. E.
(
1948
).
Origin of melanophores and their role in development of color patterns in vertebrates
.
Physiol. Rev
.
28
,
383
408
.
Rawles
,
M. E.
(
1959
).
An experimental study on the development of regional variation in the plumage pattern of the Silver Campine fowl
.
J. Morph
.
105
,
33
54
.
Rawles
,
M. E.
(
1960
).
The integumentary system
.
In Biology and Comparative Physiology of Birds. Chapter VI
, pp.
189
240
(ed.
A. J.
Marshall
).
Academic Press
.
Teillet
,
M. A.
(
1971
).
Recherches sur le mode de migration et la differentiation des nfelanoblastes cutands chez 1’embryon d’oiseau
.
Ann. Embryol. Morphog
.
4
,
95
109
.
Trinkaus
,
J. P.
(
1988
).
Directional cell movement during early development of the Teleost Blennius pholis: II. Transformation of cells of epithehal clusters into dendritic melanocytes, their dissociation from each other, and their migration to and invasion of the pectoral fin buds
.
J. exp. Zool
.
248
,
55
72
.
Twitty
,
V. C.
(
1936
).
Correlated genetic and embryological experiments on Triturus. I and II
.
J. exp. Zool
.
74
,
239
302
.
Twitty
,
V. C.
(
1937
).
The influence of nuclear factors in hybrid development studied by transplantation
.
Am. Nat
.
71
,
127
142
.
Twitty
,
V. C.
and
Bodenstein
,
D.
(
1939
).
Correlated genetic and embryological experiments on Triturus
.
J. exp. Zool
.
81
,
357
398
.
Watterson
,
R. L.
(
1942
).
The morphogenesis of down feathers with special reference to the developmental history of melanophores
.
Physiol. Zobl
.
15
,
234
259
.
Willier
,
B. H.
(
1941
).
An analysis of feather color pattern produced by grafting melanophores during embryonic development
.
Am. Nat
.
75
,
136
146
.
Wolpert
,
L.
(
1969
).
Positional information and the spatial pattern of cellular differentiation
.
J. Theor. Biol
.
25
,
1
47
.
Wolpert
,
L.
(
1981
).
Positional information and pattern formation
.
Phil. Trans. R. Soc. Lond. B
295
,
441
450
.
Wolpert
,
L.
and
Stein
,
W. D.
(
1984
).
Positional information and Pattern formation
.
In Pattern Formation: a Primer in Developmental Biology
(ed.
G. M.
Malacinski
and
S. V.
Bryant
), pp.
3
21
.
Macmillan
.