Pigment pattern expression in the plumage of the quail embryo and the quail-chick chimaera

Transcription

1 Develpment 17, (1989) Printed in Great Britain The Cmpany f Bilgists Limited Pigment pattern expressin in the plumage f the quail embry and the quail-chick chimaera MICHAEL K. RICHARDSON, AMATA HORNBRUCH and LEWIS WOLPERT Department f Anatmy and Develpmental Bilgy, University Cllege ami Middlesex Schl f Medicine, Wmdeyer Building, Cleveland Street, Lndn WJP6DB, UK Summary The pattern f pigmentatin in birds is dependent n the migratin and differentiatin f a ppulatin f neural crest cells that develp int melanblasts. On the basis f previus grafting experiments Rawles (1948) cncluded that the pigment pattern f the chimaera is determined by the gentype f the dnr melancyte. This led Wlpert (1981) t suggest that melanblasts frm ne bird can read the psitinal value f the ectderm in the feather papillae f anther bird. An alternative view is that an ismrphic prepattern in the feathers determines the pigment pattern. We have examined these ideas in relatin t the lcal pigment patterns f the embrynic quail wing, distal t the elbw, where several rws f feather papillae are cnsistently unpigmented. Melanin pigment is first seen at stage 35. By stage 39 a characteristic pigment pattern has been established. Mst f the drsal feather papillae are heavily pigmen ted, whereas many ventral papillae are unpigmented. Of the ventral papillae three rws (E2, E3 and H2) are always unpigmented, and it is these three rws that frm the basis f the quail lcal pattern. The DOPA reactin indicates that n melanblasts are present in these white feathers, althugh they are present in all the feathers f the White Leghrn wing. When quail neural crest cells are grafted t the chick, either istpically r t the wing bud, all r nearly all rws f ventral papillae becme pigmented by stage 39. The nly evidence f dnr influences in the pattern is that, in sme grafts, rws E2-3 have a high prprtin f unpigmented papillae, and wings frm earlier stages resemble the quail. When unpigmented papillae are present, histlgy shws that they cntain undifferentiated crest cells. When intrduced int a quail wing bud, chick crest cells enter all the feather papillae f the wing, including thse in rws E2-3 and H2. We suggest that neither the psitinal infrmatin nr the prepattern thery alne can accunt fr all f ur findings. Cntrary t previus claims, lcal cues may be imprtant in determining crest-cell differentiatin. We have established that crest cells migrate int all feather papillae f the quail-chick chimaera, including thse that will remain unpigmented. We shw that neither differential migratin nr differential prliferatin is invlved in pattern frmatin in the quail-chick chimaera. Key wrds: quail-chick chimaera, melancyte, psitinal infrmatin, pattern frmatin, pigmentatin, feathers, neural crest. Intrductin The pattern f pigmentatin in birds is remarkably varied and beautiful. It is nt nly f great intrinsic interest, but prvides a system fr investigating pattern frmatin and its relatin t psitinal infrmatin (Wlpert, 1969); it invlves the migratin, multiplicatin and differentiatin f just ne cell type, the neural crest cell, in a varying envirnment. The blacks, greys, brwns and rufus clurs f bird feathers, as well as the mre subdued shades f yellw, are prduced by melanin pigments synthesized by epidermal melancytes (Rawles, 196). These cells riginate in the neural crest (Drris, 1939), and fllw the drslateral migratry pathway befre entering the ectderm (Teillet, 1971; reviewed by Le Duarin, 1982). Tw theries can be put frward t accunt fr pigment-pattern frmatin, and they have very different predictins fr the pattern which ne wuld expect t see in a neural crest chimaera. Rawles (1948) grafted pigment cells between bird embrys; she cncluded that 'irrespective f the immediate surce f the melanblasts r the methd f intrducing them int freign feather germs, the results have been cnsistent in shwing that the melanphres prduce their specific clr and pattern in hmlgus feathers f varieties nrmally exhibiting an entirely different clr and pattern' (Rawles, 1948). This led Wlpert (1981) t suggest that melanblasts read the psitinal value f the feather ectderm and interpret this accrding t their wn genme (Fig. 1A). The key pint is that the psitinal values in hst and dnr are

2 86 M. K. Richardsn, A. Hrnbruch and L. Wlpert B POSITIONAL VALUE ACTIVITY f trn hctr which cntrl mrtanbfasl diftcremtliua Fig. 1. Diagram illustrating tw mdels t accunt fr pigment-pattern frmatin. (A) The psitinal infrmatin mdel. (B) The prepattera mdel. Black circles=pigmented feather papillae; pen circles=unpigmented papillae. identical, and the pattern arises frm the differential respnse f the melanblast. The predictin frm this thery is that when pigment cells are grafted between species, the pattern in the chimaera will resemble the dnr, if melanblasts can read the lcal psitinal values. A recent study f the quail-chick chimaera has prvided further evidence in favur f this thery (Kinutani and Le Duarin, 1985) even thugh the chimaera is interspecific. Yet mre evidence cmes frm grafting experiments with amphibia. Dnr-type patterns are seen in amphibian chimaerae prvided the species are nt t distantly related (DuShane, 1943; Rawles, 1948). Twitty (1936, 1937) and Twitty and Bdenstein (1939) grafted neural crest cells between different species f Triturus, and between Triturus and Ambystma. They fund that the pattern in the chimaera was determined largely by the gentype f the grafted neural crest. An alternative thery is based upn the pattern being generated by an ismrphic pre-pattem (Wlpert and Stein, 1984) in the feather papillae (Fig. IB). In this mdel there are lcal differences that determine whether r nt a feather becmes pigmented. Thus, in this case, the pattern is determined largely by the special prperties f the hst feather germs t which the melanblasts can respnd. In principle, melanblasts frm any species capable f this respnse wuld prduce a hst-type pattern when grafted t anther bird. Neither thery says anything abut the cellular r mlecular mechanisms invlved; indeed, little is knwn abut these. It culd be that differential migratin f melanblasts is invlved. Thus Wattersn (1942) and Rawles (1959) have suggested that white feathers arise because f sme blck t crest-cell entry. It is imprtant t remember that these wrkers had n reliable histlgical markers fr identifying crest cells. Alternatively, pigment cell precursrs culd migrate t all parts f the skin, but nly differentiate in certain areas. Althugh there is little evidence frm birds t supprt either view, differential migratin des seem t be imprtant in pigment patterning in embrynic and larval frms f fish (Trinkaus, 1988) and amphibia (Epperlein and Claviez, 1982). The White Leghrn has lng been used as a hst bth in pigmentatin studies (Rawles, 1948) and in studies f the quail-chick chimaera (Le Duarin, 1982). Jimbw et al. (1974) have made a detailed study f pigmentatin in this breed. Crest cells enter the feather papillae and give rise t large numbers f DOPA-psitive cells (melanblasts). These cells are present in the feather germs mainly during the perid f 9-16 days f incubatin. These melanblasts may differentiate int melanin-cntaining cells (melancytes), althugh they die befre depsiting significant amunts f melanin in the feather papilla. These shrt-lived melancytes can be seen in the basal part f feather papillae where they frm a black ring (Rawles, 1944). The defect that causes premature death in the melancytes is intrinsic t the melancytes themselves (Hamiltn, 194) and nly affects thse f the neural-crest lineage (thus the retina is pigmented in this breed because retinal melancytes arise frm the ptic cup and nt the neural crest). On the basis f genetic studies Hadley (1915) fund the White Leghrn t pssess latent pigment-patterning genes, and described this breed as 'a masquerader wh cnceals many clrs and patterns beneath her pure white plumage'. Brumbaugh (1967) cnsiders the White Leghrn t have the Extended black (E) pigmentatin gentype, a gentype it shares with such breeds as the Black Australrp, Black Minrca and Barred Plymuth Rck. While the experiments f Rawles (1948) supprt the psitinal infrmatin mdel, several crucial pints remain unreslved. It has nt been shwn that lcal patterns in the dnr are seen als in the chimaera. By lcal pattern we mean a characteristic arrangement f pigmented and unpigmented feather papillae in a defined area. The 'dnr patterns' described in the literature t date refer merely t the spts and stripes within single feathers. Our study is f grafts between the Japanese quail and the White Leghrn chick, and there are tw reasns fr this chice f breeds. First, there is n detailed descriptin in the literature f the develpment f pigment patterns in the quail-chick chimaera. Secnd, mst grafting experiments t study pigmentatin in birds have been f the chick-chick type; it is nt at all clear what happens in xengrafts. Rawles (1939) fund that rbin pigment cells grafted t the White Leghrn did nt prduce the apprpriate dnr pattern, whereas Willier (1941) fund that xengrafts using a variety f bird species all resulted in dnr clratin. Once again, hwever, difficulty arises because these authrs appeared t be lking simply at patterns and clurs within feathers, and nt at lcal patterns. Here we present a study f pigment-pattern frmatin in the quail embry, in the White Leghrn chick, and in heterspecific cmbinatins between these tw breeds.

3 Pigment-pattern frmatin in birds 87 Materials and methds Fertilised eggs f the Japanese Quail (Cturnix cturnix japnica) were btained frm a clny kept in this department. Fertilised eggs frm a strain f the White Leghrn ('Rss White') were btained cmmercially (Needle Farms, Hertfrdshire). Mapping the distributin f pigmented and unpigmented feather papillae n the wings f the nrmal quail and the perated chick Feather papillae were plucked frm the wing f the right-hand side, and the psitin f each papilla was represented n a chart with a nte f its pigmentatin type. Only distinct elevatins greater than.2 mm in height r diameter were recrded. Simple ectdermal thickenings were ignred. Three categries f pigmentatin were designated accrding t the appearance f the papilla under the dissecting micrscpe: 'pigmented', 'unpigmented' and 'trace'. A papilla is categrised as 'trace' when its melanin is visible nly at high pwer (3x) under the dissecting micrscpe. The distinctin between these categries is subjective, but they allw a large number f feather papillae t be graded in a relatively shrt space f time. Grafts f quail neural-crest cells t the chick embry Tw methds were used: (1) Istpic grafts f quail neural hemi-tubes with assciated neural crest, 2 smites in length, were made t chick embrys. Our prcedure is a mdificatin f that illustrated by Le Duarin (1982, p. 14). Dnrs and hsts were incubated fr 4-48 h at 38 C±1 C, and had between 11 and 22 smites at the time f grafting. Tissue fr grafting was excised with tungsten needles; trypsinisatin was nt used. The hemi-tubes were prduced by cutting the neural tube lngitudinally int left and right halves. The crrespnding piece f neural tube in the hst was nt divided lngitudinally, but was remved whle. The grafts were made in mst cases at the level f smites (2) Grafts f whle, quail neural tubes and assciated neural crest, 2 smites in length, were taken frm the thracic regin f a quail embry and inserted int the wing bud f a chick embry. The prcedure is essentially that f Drris (1939). The dnrs were incubated fr tw days at 38 C±1 C and had frm 14 t 18 smites at the time f grafting. The hsts were f stages (Hamburger and Hamiltn, 1951). A slit, parallel t the lng axis f the embry, was cut in the base f the right wing bud with tungsten needles and the graft wrked int the slit s that n tissue was left prtruding. With yunger hsts, the graft smetimes passed thrugh int the celm. All embrys were incubated fr a further 9-17 days befre killing. Grafting quail wing buds t the chick The right wing bud was remved frm a chick, and a quail wing bud grafted in its place. The grafted wing bud was held in place by platinum wire pins. Dnrs and hsts were f the same age within the range f stages The DOPA reactin The tissue was fixed fr 2h in 1% frmalin in phsphatebuffered saline (PBS, 4 C, ph7.4) then rinsed in PBS fr 1 h. The reactin mixture cnsisted f.1 % D-L dihydrxyphenylalanine (DOPA; Sigma) in PBS (ph7.7) which had in sme cases been gassed vernight with nitrgen. Incubatin was carried ut at 37 C fr 6-14 h during which time the reactin mixture was changed twice. Cntrls were treated as abve, but DOPA was mitted frm the reactin mixture and in sme cases 1~ 3 M-sdium diethyldithicarbamate (Sigma), a tyrsinase inhibitr, was included. Histlgy Tissue wasfixedvernight in half-strength Karnvsky's fixative (Karnvsky, 1965), dehydrated in graded alchls and embedded in Araldite. Sectins were cut at 2 /.an and stained with tluidine blue r the Feulgen-Rssenbeck technique. Results The arrangement f feather papillae in the quail and chick The maps presented here are diagrammatic rather than tpgraphic representatins f the wing surface. The feathers f the wing can be readily reslved int a series f rws rughly parallel t the prximdistal axis f the wing. These rws are arranged as discrete grups (tracts) as can be seen in Fig. 2. The basic unit f ur maps is nt the hmlgus papilla but the hmlgus rw. It was nt fund t be pssible t prduce a standard template fr cmparing hmlgus papillae n different wings. The reasn fr this is the variability in the pterylsis f different wings; nt nly des the number f papillae per rw vary frm wing t wing at a particular stage, but the number f rws in certain tracts varies als. Our maps include mst f the feather papillae distal t the elbw except thse at the wrist, and any ther papillae that are difficult t assign accurately t rws. We have given a letter and number t each rw f papillae: the letter dentes the grup (tract) t which that rw belngs, and the number indicates the psitin f the rw within the tract (Fig. 2). The mst psterir rw in each tract is number 1, the next rw number 2, and s n. The tracts cntaining the B and G rws are difficult t map accurately. These tracts are bth essentially triangular and cntain mre rws at their brad (prximal) end than at their pinted (distal) end. Additinally, B and G rws ften bifurcate r end suddenly in mid-curse. The B tract is particularly prblematic since it merges at an angle int the A rws, and the bundary is nt always easy t establish. Sme rws are highly cnsistent: rws El-4 and Hl-2 shw interesting lcal pigment patterns, and are amngst the best-defined rws f the wing. Fr these reasns, the E and H rws will frm the basis f ur reprt. Our rw designatins cmpare with the nmenclature f Lucas and Stettenheim (1972) as fllws: El and HI are under majr cverts; E2 cnsists f under median cverts; and E3-4 and H2 are under minr cverts; the 'dwns' described by these authrs are nt included in ur accunt, and must nt be cnfused with ther feathers near the psterir wing margin. Rws El and HI are cntinuus with each ther; s t are rws E3 and H2. Rws E2 and 4 have n cunterparts in the hand. Rw Fl is the mst psterir rw lying between the E and G grups. There

4 88 M. K. Richardsn, A. Hrnbruch and L. Wlpert are ther papillae in this regin but they are nt included in this accunt because they d nt always frm recgnisable rws. The I grup usually cnsists f 3 rws. If a furth is present this is disregarded s that 13 is always the lng rw n the anterir margin f the ventral hand. Fig. 3 shws light micrgraphs f typical papillae frm each categry. There is clearly sme verlap - pigmented - unplgmented * - trace f pigment Fig. 2. Quail wing (stage 37). (A and B) Ventral surface; (C and D) drsal surface. Scale bar n A=2mm. f 3A B Fig. 3. Light micrgraphs f whlemunted feather papillae illustrating the three different categries f pigmentatin. (All are frm the wing in Fig. 1.) (A) Papilla scred as 'unpigmented'; (B) scred as 'trace'; (C and D) bth scred as 'pigmented'. All are t the same scale as A. Scale bar in A=2^m.

5 Pigment-pattern frmatin in birds 89 KEY: - pigmented - unpigmented * - trace f pigment Fig. 4. Quail wing, abut stage 4. (A and B) Ventral surface; (C and D) drsal surface. Scale bar n A=5mm. 1 -i S 8 H I 6 2 pigmented El trace E1 E2 E3 E4 F1 G1 G2 G3 G4 G5 H1 H il* i E2 E3 E4 F1 G1 G2 G3 G4 G5 H1 H2 Rw between categries, althugh we have fund that mst papillae fall clearly int either the 'pigmented' r 'unpigmented' class. The nrmal pigmentatin f the quail wing We have applied chick stages t the quail; beynd stage 39, hwever, this cannt be dne accurately. Fr this reasn, stages have been gruped tgether as a single categry. Stages Traces f melanin may be seen in the incipient papillae arund the elbw as early as stage 35. Hwever, melancytes are nt seen in significant numbers until stage 36 (4 cases) n the drsal surface f the wing at the prximal ends f rws Al-4. During stages 37 (3 cases) and 38 (3 cases), pigment appears in successively mre anterir and distal papillae n the drsal surface (Fig. 2). On the ventral surface, papillae appear at Fig. 5. Histgrams illustrating the develpment f pigment in the papillae f the ventral surface f the quail wing. The abscissa shws the rw number, and the rdinate shws the mean percentage, frm all the cases f that stage, f pigmented papillae in each rw. Errr bars shw the standard errr f the mean. (A) Stage 38; (B) stage

6 81 M. K. Richardsn, A. Hrnbruch and L. Wlpert stage 37, and pigment at stage 38, at which time it appears in the newly frmed rws Gl-2, El and E4. Stages (6 cases) By this time, the definitive pattern has been established. Fig. 4 shws a quail wing f abut stage 4. All the drsal papillae are heavily pigmented with the exceptin f rw B5, and the distal members f Dl. The ventral surface has develped a characteristic pattern: rws E2-3, G5 and H2 are cnsistently unpigmented, whilst the remaining rws shw varying amunts f pigment. Since, as mentined abve, the definitin f the G rws is prblematic, it is the H and I rws that cntain the mst imprtant lcal patterns. In rws 1 -i Gl-3 and 12-3 the density f pigmentatin appraches that seen in drsal feather papillae; this gives the appearance n the ventral surface f a streak f pigmented papillae running prximdistally alng the anterir margin f the wing (Fig. 4A). Stages 38 and beynd shw a graded decrease in the frequency f pigmentatin acrss the B rws. The data are shwn in Figs 5-6, where histgrams shw the frequency f pigmented papillae in each rw. We have examined, but nt mapped, wings frm quails up t 1 mnths after hatching; the wing plumage f these birds shws bradly the same pattern as that shwn by stage embrys. Since we are t cncentrate n the E and H rws, we have examined these in a further 11 embrys, making a I pigmented trace n Tt in ID CO CO CD CD O O O O in *- O Q 1 1 Rw Fig. 6. Histgrams illustrating the develpment f pigment n the drsal surface f the quail wing. (A) Stage 36; (B) stage 37; (C) stage 38; (D) stage

7 Pigment-pattern frmatin in birds 811 ttal f 17 cases f stages Nne shwed any pigment in E2-3 r H2. Feathers in El and HI are black r grey at the base nly, r base and tip nly, with white intervening. One third f cases shw n pigment at all in E4. The intensity f pigment in a rw generally declines twards its distal end. Black nly, and nt brwn, is seen in the E and H rws. The nly exceptins are the ccasinal embrys whse entire plumage lacks black melanin; in these cases, brwn replaces black in all feathers. 4 > E3 7A X Fig. 7. Cleared whle munts f nrmal White Leghrn feather papillae. (A) Papilla frm rw E3 (stage 39). Melancytes are seen in the skin, the shaft f the papilla, and in a ring at the base f the papilla. Scale bar=25//m. (B) Skin frm stage 42 wing shwing the prminent melancyte rings (arrw). Scale bar=25/an. Fig. 8. Cleared whle munt f White Leghrn skin (11 days incubatin) after treatment with DOPA. Scale bar=4jfln. - pigmcntcd - unpigmcntcd * = trace f pigment Fig. 9. Chick wing (stage 37) after istpic graft f quail neural crest. (A and B) Ventral; (C and D) drsal. Scale bar n A=4mm.

8 812 M. K. Richardsn, A. Hrnbruch and L. Wlpert 14 quail wings,, frm stages 36-41, were examined with the DOPA reactin. The ther wing frm each embry was untreated. A psitive reactin appeared t be cnfined t thse papillae that were already in the prcess f becming pigmented; thus there was little difference between treated and untreated wings. In n cases were any DOPA-psitive cells fund in the unpigmented rws (E2-3, H2) f the ventral surface. Pigmentatin f the White Leghrn The plumage appears entirely white t the naked eye, althugh punctate r fusifrm melancytes can be seen under the micrscpe in cleared specimens. We have examined 15 wings frm stages Melancytes are seen first in the larger papillae n the drsal surface, and in the skin f the cubital apterium, at the end f stage 37. Of the 9 wings examined f stages 37-39, 6 had n melancytes at all, but by stage 42, all wings shwed melancytes at the bases f all their papillae. Fig. 7A shws a papilla (rw E3) frm a nrmal stage 39 White Leghrn. Immature punctate melancytes are seen in the skin, in a ring at the base f the papilla, and scattered up the shaft f the papilla. All the feathers f the wing appear t be capable f supprting the differentiatin f these cells. By stage 42 (Fig. 7B), the rings have becme particularly prminent. The DOPA reactin reveals large numbers f melanblasts in White Leghrn feather papillae (cf. Jimbw et al. 1974). Mrever, they are present in all the feather papillae f the wing, including El-4 (Fig. 8) and Hl-2. These DOPA-psitive cells are nt seen if diethyldithicarbamate, an inhibitr f tyrsinase, is present in the reactin medium. This indicates that ur technique is specific fr melanblasts. Pigmentatin patterns f the wing f chicks after quail neural tube grafts Istpic grafts Of 91 grafts perfrmed, 45 survived and 27 f these were either entirely unpigmented, r shwed irregular patches f pigment, usually in the feathers f the drsal frearm. In these wings, the ventral frearm and at least the distal half f the hand are unpigmented. These cases are nt included in ur accunt since we cnsider that they represent a failure f the grafted quail cells t becme prperly incrprated int the hst. This leaves 18 cases. The right wing was chsen fr study wherever pssible but, in sme cases, nly the left wing had becme pigmented and was taken instead. In either = pigmented - unpigmented * - trace f pigment Fig. 1. Chick wing (stage 4) after istpic graft f quail neural crest. (A and B) Ventral; (C and D) drsal. Scale bar n A=7mm.

9 Pigment-pattern frmatin in birds pigmented trace E1 E2 E3 E4 F1 G1 G2 G3 G4 G5 H1 H E1 E2 E3 E4 F1 G1 G2 G3 G4 G5 H1 H E1 E2 E3 E4 F1 G1 G2 G3 G4 G5 H1 H Rw Fig. 11. Histgrams illustrating the develpment f pigment in the istpic graft, ventral surface. (A) Stage 37; (B) stage 38; (C) stage case, nly ne wing was taken frm a single embry. In all f the cases examined, pigment spread beynd the axial level f the gtaft, indicating that the number f grafted cells was sufficient t ppulate the entire wing regin. Stages 36 (3 cases) and 37 (6 cases) resemble the quail althugh pigmentatin appears t develp mre rapidly than in quails f the same stage. A stage 37 graft is shwn in Fig. 9. At stage 38 (3 cases), there appears t be a preccius expressin f pigment, cmpared t the nrmal quail, n the ventral wing surface. During stages 38-41, quite a different pattern emerges whereby mst f the feathers f the wing becme pigmented. In wings f stages 39-41, (6 cases) virtually all f the drsal papillae are pigmented and mst f thse n the ventral surface are t (Fig. 1). Nearly all the rws f the wing cntain a high prprtin f pigmented papillae: at least 8% in mst cases (Figs 11, 12). Rws E2-3 are exceptinal in that fewer than half f their papillae are pigmented (the mean percentage f pigmented papillae in these rws is: E2=36%, E3=43%). These data are the means frm several cases. Rw H2 is cnsistently unpigmented in the quail, but in the perated chick wings it is always pigmented. T give sme idea f the variability between grafts, the E and H rws were examined in a further 5 grafts making a ttal f 11 grafts f this stage. Rws E2-3 were entirely pigmented in 4 cases, entirely unpigmented in 2 cases, and shwed a mixture f pigmented and unpigmented papillae in 5 cases. Rw H2 always cntained pigmented papillae. Fr histlgical analysis, 44 grafts were perfrmed prviding 29 cases. The early phase f crest-cell migratin in the chimaera, including the seeding f the ectderm, has been cnsidered by Teillet (1971). We cnsider stage 33 nwards. Transverse sectins f the frearm shw quail cells in the ectderm, the lse cnnective tissue beneath the dermis and in neurvascular bundles. Frm stage 37 nwards, a netwrk f pigmented cells is frmed which extends rund the full circumference f the limb under the dermis, and between the blcks f skeletal muscle f the limb. Quail cells were at all stages rare in the dermis, and were never seen in cartilage r in the dermal pulp f feather papillae. Quail cells became distributed t the ectderm rund the entire circumference f the limb. We culd find n evidence t supprt the idea that crest cells migrate mre readily int pigmented papillae than int unpigmented papillae, r that they prliferate mre in pigmented papillae. In 4 wings f stage 39, the prprtin f ectdermal cells bearing the quail nuclelar marker was determined. In the heavily pigmented basal part f El papillae, this prprtin was in the range 1/1243-1/296, and all f the quail cells were pigmented. Melancytes were always lcated in their definitive psitin at the apex f the barb ridges. The characteristic quail nuclelus becmes dispersed and faint as the melancyte matures. Therefre, t see if we were underestimating the number f quail cells, we did sme mre cunts scring nt fr the nuclelar marker, but simply fr melancytes. Our criteria were that the nucleus be cmpletely surrunded by melanin granules, but nt verlaid by them. This prduced values f the same rder f magnitude, the range being 1/442-1/32. Next, papillae frm rws E2-3 were examined. They were either entirely unpigmented, r had nly slight traces f pigment at their bases. Quail cells, always unpigmented, were seen in all the papillae and the frequency was in the range 1/468-1/432. The mean (1/666) is in the range f the frequency f quail cells seen in pigmented papillae. The unpigmented quail cells were always lcated in the definitive melancyte psitin at the apex f the barb ridges (Fig. 13). Grafts f quail neural tube t the chick wing bud 22 grafts were perfrmed, all f which survived, prviding 8 cases f stages fr examinatin. The remainder were nly partially pigmented and are nt included in ur analysis. The wing shwn in Fig. 14 has a similar pattern t that f the istpic graft at the same stage (cf. Fig. 1). A cmparisn between Figs 11C and 15A shws that in bth cases all ventral rws, including

10 814 M. K. Richardsn, A. Hrnbruch and L. Wlpert pigmented trace CO < < ID < < < < CD < < CO eg CO CO CO C\J O CO O in *- O Q 1 < < < Fig. 12. Histgrams illustrating the develpment f pigment in the istpic graft, drsal surface. (A) Stage 36; (B) stage 37; (C) stage 38; (D) stage rws E2-3 and H2, cntain pigmented papillae (E2=3%, E3=23% and H2=86%). As was the case with istpic grafts, the mean frequency f pigmented papillae in rws E2-3 is lwer than in all f the ther rws. Thus a similar pattern is prduced by quite a different grafting technique. Grafts f quail wing buds t the White Leghrn chick 28 grafts were perfrmed, 5 f which died leaving 23 cases. In 16 cases, the dnr quail was f stage 16-2, and 15 f these develped int cmpletely unpigmented wings. One case (dnr stage 18) had a few greyish papillae n the drsal frearm but was therwise unpigmented. Of 5 cases that used stage 21 dnrs, 2 develped int unpigmented wings, and 3 int pigmented wings. A stage 22 wing bud, and stage 25 wing tip, bth develped int pigmented wings. Thus it seems that neural crest cells first enter the wing f the nrmal quail at stage 21 r later. Unpigmented quail wings frm this series were treated with the DOPA reactin. Of 12 cases treated (stages 38-39), 4 shwed melanblasts in all the feather germs f the wing, including rws El-4 and Hl-2 (Fig. 16). Histlgy, cupled with Feulgen-Rssenbeck staining, cnfirmed the view that these cells were chick melanblasts that had migrated int the quail wing. Five wings shwed melanblasts distributed in irregular patches, as thugh the chick cells had failed t enter the wing in sufficient

11 Pigment-pattern frmatin in birds A Fig. 13. Light micrgraph shwing a transverse sectin thrugh an unpigmented papilla (rw E3, istpic graft, stage 39). (A) Phase cntrast. Scale bar=2^m. (B) Inset shwn in 13A, bright-field. Nte the unpigmented quail cell at the apex f the barb ridge (arrw). Scale bar=1/an. - pigmented - unpigmented * - trace f pigment Fig. 14. Chick wing (stage 39-4) after a graft f quail neural crest int the wing bud. (A and B) Ventral; (C and D) drsal. Scale bar n A=7 mm. numbers. A further three wings f this type had diethyldithicarbamate included in the DOPA reactin mixture. These shwed n staining at all, which suggests that the reactin is specific fr melanblasts. Discussin The nrmal quail embry shws a characteristic pigment pattern in its wing plumage. The mst striking features f the pattern are: (1) the cnsistently heavy pigmentatin f the drsal surface cmpared t the ventral; (2) the graded decrease in pigment acrss rws B1-B5; (3) the dark streak f feathers near the anterir margin f the ventral surface, frmed by the heavy pigmentatin f rws Gl-3 and 12-3; and (4) the cnsistent lack f pigment in rws E2-3 and H2. Features 1-3 are dependent n the relative amunt f melanin in individual papillae; feature 4, hwever, the predictable absence f pigment in a given set f feather papillae, defines a lcal pattern.

12 816 M. K. Richardsn, A. Hrnbruch and L. Wlpert pgmented trace E1 E2 E3 E4 F1 G1 G2 G3 G4 GS H1 H < < < < < < < < < ' - m m c j 3 r m c D r j c Q m c D t D < Rw We find that when quail neural crest is grafted istpically t the chick, the pigment patterns prduced in the chick feathers at stages resemble thse f the nrmal quail, althugh pigmentatin develps smewhat mre rapidly than in the quail. But by stages 39-41, the pattern in the hst has becme very different frm the pattern f the nrmal quail wing. Althugh the drsal surface is heavily pigmented like the quail, the in T- Fig. 15. Histgrams illustrating the develpment f pigment in the chick wing fllwing a graft f quail crest t the chick wing bud. (A and B) bth=stage ventral surface, instead f shwing well-defined lcal patterns, shws pigment in mst r all f its rws. In ther wrds, the chick hst shws a generalised increase in the frequency f pigmented papillae cmpared with the nrmal quail. Grafting quail neural crest int the chick wing bud prduces in the chick plumage, at stages 39-41, the same pigmentatin as is prduced by an istpic graft; that is, melanin is fund in mst r all Fig. 16. Unpigmented quail wings prduced by grafting quail wing buds befre stage 21 t the White Leghrn. (A) Treated with DOPA and diethyldithicarbamate (gassed with N 2 ), stage 38. Scale bar=2mm. (B) Treated with DOPA nly (stage 38). Scale bar=2mm. (C) Cleared whlemunt frm B (bxed area) shwing chick melanblasts in the papillae f the E grup. L

13 Pigment-pattern frmatin in birds 817 f the rws f papillae. This shws that the methd f grafting des nt influence the result. Althugh the pigment pattern f the chimaera is very different frm that f the dnr, there is sme indicatin f dnr pattern in tw rws f papillae, E2 and E3. It will be remembered that in the quail, these tw rws are always unpigmented. After bth types f gtaft, mst f the rws n the hst wing at stages shw pigmentatin in mre than 8% f their papillae. The exceptins are rws E2 and E3 in which, n average, fewer than half f the papillae are pigmented. Indeed, in sme cases, pigment is entirely absent frm these rws. If these unpigmented papillae are examined histlgically, undifferentiated quail cells are fund t be present. Anther rw that is always unpigmented in the quail is rw H2; in perated chicks hwever, this rw is always pigmented. It is pssible that quail crest cells can read the psitinal value f rws E2-3 in the chick. Kinutani and Le Durain (1985) have reprted that the drsal stripes n the quail embry are seen als in the chimaera, and have interpreted this as indicating that quail cells can read the psitinal values f chick feathers. Our studies f the White Leghrn have shwn that all the feather germs f the wing are capable f supprting at least the initial stages in the differentiatin f melancytes. Our evidence fr this is f tw types. First, rings f melancytes are fund at the bases f all the feather papillae f the White Leghrn wing. Secnd, DOPA-psitive melancytes are present in all the feather germs f the wing. In the nrmal quail, by cntrast, nly thse feathers that becme pigmented cntain DOPA-psitive melanblasts; the unpigmented papillae n the ventral surface d nt. By grafting quail wing buds f different stages t the chick, we find that crest cells d nt enter the quail wing bud in significant numbers until stage 21 and later. Our results d nt supprt the hypthesis f Rawles (1959) and Wattersn (1942) which says that white feathers arise because f sme physical blck t melanblast entry. Our reasning is based n tw lines f evidence. Our histlgical analysis f the quail-chick chimaera shws that undifferentiated neural crest cells are present in white feather papillae, and their numbers in these papillae are cmparable t the number f quail cells in pigmented papillae. Furthermre, we find that when quail wing buds are grafted t the chick, DOPA staining reveals that chick melanblasts have migrated int all the quail feathers, including thse in rws E2-3 and H2 which are, in the nrmal quail, unpigmented. Our results cannt be explained by either the psitinal infrmatin r prepattem theries alne. The lw frequency f pigmented papillae in E2-3 in the chimaera suggests that quail cells can read the psitinal values f the ectderm in thse papillae. Hwever, this des nt apply t the ther papillae f the chick wing; mst hst papillae becme pigmented, including thse in H2 that are always unpigmented in the quail. In these, the quail cells may be respnding t lcal cues in the papillae; we have shwn that all the papillae f the chick wing are capable f supprting at least the early stages f chick melancyte differentiatin. When chick crest cells are intrduced int the quail wing, they enter all rws f papillae and differentiate int melanblasts. This means that either there is n prepattem in the quail wing r, if there is ne, chick melanblasts d nt respnd t it. We prpse the fllwing mdel t accunt fr pigment-pattern frmatin in the plumage f the avian embry wing. Neural crest cells enter the wing bud at stage 21 where they subsequently becme distributed t all regins f the ectderm. Crest cells enter all feather papillae, including thse destined t frm white feathers, and they cme t ccupy the apical regin f the barb ridges. Lcal cues in the feather papillae cntrl crest cell differentiatin such that, in pigmented feathers, dendritic melancytes are frmed whereas, in white feather papillae, the crest cells remain undifferentiated with n tyrsinase activity. Whether these lcal cues represent a respnse f the crest cell t psitinal infrmatin, r t special lcal prperties cnstituting a prepattem remains t be determined. The White Leghrn has been the breed f chice as a hst in pigmentatin studies fr many years. It culd, hwever, be argued that the lack f pigmentatin in this breed leads t anmalus interactins. We are therefre investigating pigmentatin patterns in chimaerae cnstructed frm ther bird species. Preliminary findings suggest that a prepattem specified by psitinal infrmatin can best accunt fr mst f the results. We are grateful t Dr Cheryll Tickle fr her cmments n the manuscript. This wrk is supprted by a grant frm the Science and Engineering Research Cuncil. References BRUMBAUGH, J. A. (1967). Differentiatin f black-red melanin in the fwl: interactin f pattern genes and feather fllicle milieu. J. exp. Zl. 166, DORRIS, F. (1939). The prductin f pigment by chick neural crest in grafts t the 3-day limb bud. J. exp. Zl. 8, DUSHANE, G. P. (1943). The embrylgy f vertebrate pigment cells. Part 1. amphibia. Q. Rev. Bil. 18, EPPERLEIN, H. H. AND CLAVIEZ, M. (1982). Frmatin f pigment cell patterns in Triturus alpestris embrys. Devi Bil. 91, HADLEY, P. B. (1915). The White Leghrn - a masquerader wh cnceals many clrs and patterns under her pure white plumage. J. Hered. 6, HAMBURGER, V. AND HAMILTON, H. L. (1951). A series f nrmal stages in the develpment f the chick embry. J. Mrph. 88, HAMILTON, H. L. (194). A study f the physilgical prperties f melanphres with special reference t their rle in feather clratin. Anat. Rec. 78, JlMBOW, K., SZABO, G. AND FlTZPATRICK, T. B. (1974). Ultrastructural investigatin f autphagcytsis f melansmes and prgrammed death f melancytes in White Leghrn feathers: a study f mrphgenetic events leading t hypmelansis. Devi Bil. 36, KARNOVSKY, M. J. (1965). A frmaldehyde glutaraldehyde fixative f high smlanty fr use in electrn micrscpy. J. Cell Bil. 27, 137a. (Abstr.) KINUTANI, M. AND LE DOUAJUN, N. M. (1985). Avian spinal crd chimeras. 1. Hatching ability and psthatching survival in hmand heterspecific chimeras. Devi Bil. Ill,

14 818 M. K. Richardsn, A. Hrnbruch and L. Wlpert LE DOUARIN, N. M. (1982). The Neural Crest. Lndn/New Yrk: Cambridge Univ. Press. LUCAS, A. M. AND STETTENHEIM, P. R. (1972). In Avian Anatmy: Integument, Part 1. Washingtn D.C.: United States Department f Agriculture. RAWLES, M. E. (1939). The prductin f rbin pigment in white leghrn feathers by grafts f embrynic rbin tissue. /. Genet. 38, RAWLES, M. E. (1944). The migratin f melanblasts after hatching int pigment-free skin gtafts f the cmmn fwl. Physil. zl. 17, RAWLES, M. E. (1948). Origin f melanphres and their rle in develpment f clr patterns in vertebrates. Physil. Rev. 28, RAWLES, M. E. (1959). An experimental study n the develpment f reginal variatin in the plumage pattern f the Silver Campine fwl. J. Mrph. 15, RAWLES, M. E. (196). The integumentary system. In Bilgy and Cmparative Physilgy f Birds. Chapter VI, pp (ed. A. J. Marshall). Academic Press. TEJLLET, M. A. (1971). Recherches sur le mde de migratin et la diffe'renciatin des me'lanblastes cutane's chez l'embryn d'iseau. Ann. Embryl. Mrphg. 4, TRINKAUS, J. P. (1988). Directinal cell mvement during early develpment f the Telest Blennius phlis: II. Transfrmatin f cells f epithelial clusters int dendritic melancytes, their dissciatin frm each ther, and their migratin t and invasin f the pectral fin buds. J. exp. Zl. 248, TWITTY, V. C. (1936). Crrelated genetic and embrylgical experiments n Triturus. I and II. /. exp. Zl. 74, TWITTY, V. C. (1937). The influence f nuclear factrs in hybrid develpment studied by transplantatin. Am. Nat. 71, TWITTY, V. C. AND BODENSTEIN, D. (1939). Crrelated genetic and embrylgical experiments n Triturus. J. exp. Zl. 81, WATTERSON, R. L. (1942). The mrphgenesis f dwn feathers with special reference t the develpmental histry f melanphres. Physil. Z6l. 15, WILUER, B. H. (1941). An analysis f feather clr pattern prduced by grafting melanphres during embrynic develpment. Am. Nat. 75, WOLPERT, L. (1969). Psitinal infrmatin and the spatial pattern f cellular differentiatin. 1. Ther. Bil. 25, WOLPERT, L. (1981). Psitinal infrmatin and pattern frmatin. Phil. Trans. R. Sc. Lnd. B 295, WOLPERT, L. AND STEIN, W. D. (1984). Psitinal infrmatin and Pattern frmatin. In Pattern Frmatin: a Primer in Develpmental Bilgy (ed. G. M. Malacinski and S. V. Bryant), pp Macmillan. (Accepted 21 September 1989)