J. Embryol. exp. Morph.

74, 245-259 (1983) 245

Printed in Great Britain © The Company of Biologists Limited 1983

The mechanism of feather pattern development in

the chick
I. The time of determination of feather position
By DUNCAN DAVIDSON 1
From the MRC Clinical and Population Cytogenetics Unit, Western General
Hospital, Edinburgh

SUMMARY
A regular array of feather primordia covers chick dorsal skin in vivo. The pattern develops
over a period of 2 days as a morphogenetic wave sweeps across either side of the back of the
chick, forming successive anteroposterior rows of primordia.
This paper describes a new method for the culture of chick skin which allows the develop-
ment of large areas of the feather pattern to be investigated experimentally. Skin is cultured
on a substratum of hydrated collagen; since the collagen is transparent, feather primordium
development can be observed in detail.
The new method has been used to investigate the problem of when the positions of feathers
are determined. I show that during the time when thefirstfew rows of primordia are forming,
skin taken from just lateral to the most recently formed row can be caused to form an increased
number of primordia per row by stretching it anteroposteriorly. This result indicates that the
positions of feathers are determined sequentially along an invisible wave which moves just
ahead of the visible wave of primordium morphogenesis.

INTRODUCTION
Feather primordia develop over well-defined regions of chicken skin
(pterylae) in an exquisitely regular pattern (Fig. 1A). The primordia in each
pteryla do not develop synchronously. In the posterior part of the spinal pteryla,
for example, an initial row of feather primordia forms along the dorsal midline
on the 6th day of incubation (stage 29; Hamburger & Hamilton (1951)) and, over
the next 2 days, about nine successive rows form on either side of this initial row
(see Sengel (1976) for a review). The development of the pattern is, therefore,
a strikingly dynamic process. A morphogenetic 'wave' (Zeeman, 1974) sweeps
across the skin: ahead of the wave, the tissue lacks any obvious pattern; behind
it, primordia have formed in an orderly triangular array.
The movement of the visible, morphogenetic wave suggests the movement of
a prior, invisible wave of determination at which cells become committed to form
either primordia or interplumar skin. This is an attractive possibility because it

1
Author's address: MRC Clinical and Population Cytogenetics Unit, Western General
Hospital, Crewe Road, Edinburgh, EH4 2XU, U.K.
246 D. DAVIDSON
suggests that the pattern could be generated by a template mechanism by which
the disposition of feather sites in one row governs the positions of sites in the
next. Published models show how such a mechanism could generate the trian-
gular array (Ede, 1972; Novel, 1973; Sengel, 1976).
Published evidence does not, however, provide an unambiguous demonstra-
tion that feather positions are determined row by row. Linsenmayer (1972a,b)
and McLachlan (1980) have shown that feather positions in the thigh and limb
are established less than about 1 day before primordia form. However, the only
experimental evidence for row by row determination comes from in vitro experi-
ments in which the pattern of extant primordia disappears following explanation
and a new pattern is formed sequentially, beginning where a row was just starting
to form in the original one (Novel, 1973). These experiments were unable,
however, to establish that the new pattern formed de novo rather than by
reformation of primordia at their original positions (Novel (1973) page 628), so
that the relevance of this result to the normal mechanism of pattern formation
is not clear. The problem of precisely when the positions of individual feathers
are determined therefore remains to be solved.
The present paper describes a new method for culturing skin and reports an
in vitro experimental analysis of this problem. The results show that feather
positions are determined row by row, a few hours ahead of primordium forma-
tion.

MATERIALS AND METHODS

Organ culture
Dorsal skin was cultured on an island of hydrated collagen surrounded by
liquid tissue-culture medium.

Preparation of organ-culture dishes

Collagen substrata were made by raising the pH and ionic strength of a sterile
solution of rat-tail tendon collagen in very dilute HC1 (pH4-0) to physiological
levels using 10 x strength modified Eagles medium (Flow Laboratories) and
0-142M-sodium hydroxide (Elsdale & Bard, 1972). Newborn calf serum (10 %)
and ascorbic acid (50/ig/ml) were incorporated in the mixture before setting.
(Preliminary experiments showed that ascorbic acid reduced the extent of cell
outgrowth from skin explants and thus helped to maintain the integrity of the
skin.) About lml of the collagen mixture was allowed to set within a sterile,
2 cm-diameter fence of wire gauze (stainless steel, 60 divisions per inch, United
Wire, Edinburgh) in the centre of a 5 cm Petri dish (NUNC). 2ml of culture
medium was added to surround this collagen island. The culture medium com-
prised F10 (Flow Laboratories), supplemented with newborn calf serum (10 % ) ,
glutamine (146-5 mg/ml), ascorbic acid (50 /ig/ml), penicillin (100 units/ml) and
streptomycin (100/ig/ml), buffered to pH7-2 with MOPS buffer.
 Time of determination of feather position in chick 247

Skin preparation
White-Leghorn or Brown-Leghorn eggs (from the Poultry Research Unit,
Roslin, Edinburgh) were incubated at 38-5 °C in a humidified atmosphere. At
the appropriate stage of development, embryos were decapitated and trans-
ferred to a sterile Petri dish containing solidified paraffin wax, to a depth of about
2 mm, and sufficient saline (Dulbecco A, Oxoid) to submerge the embryos: each
embryo was pinned to the wax through shoulders and tail. Using fine scissors and
forceps, a rectangle of skin encompassing the entire width of the prospective
pteryla and extending on either side to the thigh pteryla was gently excised and
placed flat on a sterile glass coverslip. The coverslip was removed from the saline
and its tip applied to the surface of the collagen substratum in an organ-culture
dish so that the skin slipped with its dermis downward onto the collagen with
minimum distortion. The number of rows of primordia was counted immediately
after explanation. The culture was then incubated at 38-5 °C in a humidified,
well-ventilated atmosphere.

Operative treatments
Cutting: skin was laid on a sterile glass coverslip in a Petri dish and cut using a
sterile scalpel: each blade was used for only two or three cuts and was dried
between cutting to avoid displacing the skin by surface tension.

Stretching: skin was dragged gently over the dry surface of the collagen sub-
stratum in an organ-culture dish to which no culture medium had been added.
The dish was then tilted for a few minutes to drain from the surface of the
collagen any saline which had been carried over with the skin. The skin was then
allowed to adhere to the substratum by incubating the dish for 1 h before adding
medium. The increase in length obtained by this method was variable. In about
two thirds of the specimens treated, an adequate increase in length for the
purpose of the experiment (more than 20% longer than the contralateral
control) was achieved and maintained. Specimens which fell below this standard
were discarded.

Addition of colcemid: colcemid was added to the medium surrounding the

collagen. The medium was mixed and washed several times over the explant.

Observation of skin development

Living skin was examined under a Wild dissecting microscope, usually at x 16
magnification, using bright-field transmitted illumination. Measurements were
made at x 16 using a Wild micrometer eyepiece (12mm; 120 divisions). Optimal
detection of primordia in the early stages of their development was achieved by
tilting the dish to an angle of about 50° under the microscope. The numbers of
248 D. DAVIDSON
rows of primordia referred to in the Results indicate all the rows from one side
of the pteryla to the other.
Skin for histological examination was embedded in paraffin wax for serial
sectioning, and in methacylate for optimal morphology of single sections. Skin
was floated onto a millipore filter material and the excess liquid drained off so
that the skin attached to the filter and remained flat during processing. This
assembly was fixed in 2-5% glutaraldehyde (buffered with cacodylate (0-1 M,
100 mOsM) to pH 7-4) and dehydrated through graded alcohols. For wax embed-
ding, the assembly was cleared in xylene, and embedded in the usual manner.
Sections 7/im thick were cut with a Leitz microtome and stained with
haematoxylin and eosin. For methacrylate embedding, a modification of the
method of Lee (1977) was used. Skin was removed from the filter after treatment
in absolute alcohol, impregnated with Solution A (2-butoxy-ethanol, 8 ml; ben-
zol peroxide, 0-5 g; 2-hydroxyethyl methacrylate, 80ml) through two 24h
changes and embedded in methacrylate (Solution A, 42 ml; polyethylene glycol,
8 ml; n.n. dimethylanaline, lml). Sections 5pim thick were cut using a Ralph
glass knife (made on an LKB histo knifemaker) in a Leitz microtome and stained
with solochrome yellow. Sections were examined under a Zeiss Universal
microscope using bright-field illumination and, where appropriate, Nomarski
optics.

RESULTS
Preliminary observations on the development of skin in culture
Since the present work employs a new culture method, a brief description of
development in vitro is necessary before dealing with the experimental evidence.
Cultured skin from embryos older than 5 days 18h of incubation, including
prospective pterylae with no visible signs of feather development, reliably
formed a normal pattern of primordia (Fig. 1A and IB) which, over a period of
5 to 10 days, formed feather filaments (Sengel, 1971) a few millimeters long.
The only major departure in culture from the normal spatial development of
the pattern as a whole was that cultured skin did not expand (compare Fig. 1A
and IB). Indeed, the skin shrank by about 10 % in linear dimensions during the
first 2 h in culture. In normal development the width of the pteryla increases by
a factor of about two while primordia are forming, and a large pteryla, containing
17 to 21 rows, is produced. Since no such expansion occurred in culture, skin
explanted when the first row of primordia was forming produced a small pteryla
with only about nine rows. When older, and therefore larger, pterylae (contain-
ing up to five rows) were explanted and cultured, the number of normally spaced
rows which eventually formed varied in proportion to the size of the pteryla.
Minor abnormalities in the spatial development of the feather pattern com-
prised areas where ten or more primordia, although roughly equidistant, were
distributed in an irregular manner (in about 20 % of several hundred specimens),
 Time of determination of feather position in chick 249

\\i\ B ••••-•. ., •"-

* / v .. • ' : ••

Fig. 1. Comparison of feather patterns formed in vivo and in culture.

The figure shows the lumbosacral region of the spinal pteryla in freshly excised
skin (A), and in cultured skin (B) which was explanted when three rows of primordia
were visible and allowed to develop for 2 days in culture. In both specimens the
development of the pattern is nearly complete. 18 rows are present in (A) and 11
rows in (B). Magnification of (A) and (B): xl2. Scale bar: lmm. Arrow: pteryla
boundary, t: part of the thigh pteryla.

and small areas where two feather primordia were joined as if they had been
doubled or fused (in about 30 % of specimens). These 'fused' primordia were
similar in appearance to the more extensively fused primordia and ridges which
have been described in embryos from a scaleless mutant line selected for feather
production (Brotman, 1976). None of these abnormalities was observed in
several hundred specimens of freshly excised skin, although additions or
deletions of chevron rows and displacements of single primordia (Linsenmayer,
1972a; Harris, 1972), were common.
The temporal development of the feather pattern in cultured skin differed
markedly from development in vivo. Successive rows formed at intervals of
about 6 h in vivo and about 15 h in skin explanted at stage 29 (Davidson, 1983).
Histology of specimens prepared in parallel from cultured, and freshly ex-
cised, skin showed that feather primordia which formed in culture comprised
well-defined epidermal placodes and dermal condensations similar to those for-
med in vivo (Fig. 2A-D). However, in cultured, as compared with freshly ex-
cised, skin primordia in the two most recently formed rows were flatter (compare
Fig. 2D with Fig. 3A), epidermal placodes were thinner, and epidermal spurs
and anchor filaments (Wessells, 1965; Kischer & Keeter, 1971) were much less
common; the dermal cell population density was clearly abnormally high, the
individual dermal cells were generally short and lacked the bipolar morphology
characteristic of cells in freshly excised skin, and the well-known cellular arrays
which link adjacent condensations in vivo (Stuart & Moscona, 1967) were poorly
250 D. DAVIDSON

* »' • \ *• V-

Fig. 2
 Time of determination of feather position in chick 251

Table 1. Percent mitoses in the dermis of cultured skin

Region Number of cells counted Percent mitoses
Intercondensation 2974 0-91
Condensation 11806 1-46

developed and in some cases absent (compare Fig. 2A with Fig. 2B). The cells
within condensations in cultured skin were, however, tightly packed and con-
tained enlarged nuclei similar to those present in condensations formed in vivo
(Wessells, 1965).
The frequency of mitoses in untreated, cultured skin (Table 1) showed that
proliferation continued both in the intercondensation dermis and in the con-
densations of the two most recently formed rows on either side of the pteryla.
Dividing cells were also common in the epidermal placodes. In order to examine
more closely the frequency of mitoses in the dermis, specimens which had been
allowed to develop in culture for 2 days were treated with colcemid (final con-
centration 5-4 x 10~6 M) for 5-5 h prior to fixation. Three well-defined condensa-
tions in the most recently formed row were selected and the frequency of mitoses
and the position of each dividing cell were assayed in six consecutive transverse
sections through the central region of each condensation. In each case, the
frequency of mitoses was roughly uniform across the diameter of the condensa-
tion and the surrounding dermis (Fig. 4).

Histological corroboration of observations made on living skin

In the following experiments it is important to know if primordia are already
becoming histologically differentiated in advance of those which are visible
under the dissecting microscope. This point was examined by counting the rows
of primordia in 14 specimens of freshly excised, and 11 specimens of cultured,
skin first in the dissecting microscope and then in serial transverse or tangential
sections. (In sections, structures were scored as primordia if they comprised both
an epidermal placode and a clearly discernible, though not necessarily fully
developed, dermal condensation.)

Fig. 2. The histological structure of feather primordia formed in cultured skin.

(A, B, C): Frontal sections. (D): Transverse section.
For the purpose of comparison, a section through a developing condensation in
freshly excised skin is shown in (A): this primordium is in the row adjacent to the
most recently formed row. (B) shows a condensation which developed in culture: this
primordium is also in the row adjacent to the most recently formed row. (C) shows
a mature condensation formed in culture. (D) shows a primordium in the most
recently formed row in cultured skin.
Magnification of (A-D): X230. Scale bar: 100 ^m. Arrow: an array of polarised
dermal cells aligned towards an adjacent condensation in (A); compare with the
relatively poor organisation of arrays in (B).
252 D. DAVIDSON
The results confirm that the formation of the feather pattern as observed under
the dissecting microscope corresponds row for row with its histological develop-
ment. Where less than nine rows of primordia were present, the number of rows
detected by either method was the same (to within one row); neither method
consistently revealed more rows than the other. In more-advanced specimens,
it proved difficult to establish unequivocally the number of rows from serial
sections. It was clear, however, that a well-defined region of histologically unpat-
terned skin (one to two rows wide in pterylae with 14 to 16 rows of primordia)
lay between the most lateral row and the edge of the prospective pteryla where

Fig. 3. The structure of skin immediately lateral to the most recently formed row of
primordia.
(A and B): Freshly excised skin. (A): Transverse section (detail) of a specimen
with seven rows of primordia. (B): Frontal section (detail) of a specimen with five
rows of primordia visible in vitro. (In each of several sections of (B) all five rows were
visible and the plane of section passed through equivalent rows on either side of the
dorsal midline at approximately the same level. This indicates that the material was
sectioned close to the plane of the epidermis and eliminates the possibility that the
right part of the area in the figure undercuts well-developed primordia.)
Note that the formation of primordia is quite abrupt, there being only one row of
structures (arrow) intermediate between morphologically unpatterned skin (u) and
well-formed primordia (/?). Magnification: X140. Scale bar: 100/zm.
 Time of determination of feather position in chick 253
dense dermis and columnar epidermis gave way to loose mesenchyme and squat
epithelial cells.
None of the specimens examined showed histological evidence of patterned
morphogenesis in the epidermis or dermis more than one row lateral to primor-
dia comprising both a placode and a condensation (Fig. 3A and 3B).

The time of determination of feather positions: the development of stretched skin

To find out when the positions of feathers are determined, it is necessary to
measure the commitment of skin to form primordia at particular places. The
measure used in the following experiment was to stretch the skin along the length
of a prospective feather row and to ask if the skin, though now longer, remains
committed to form the normal number of primordia per row.
Developing pteryla with between 0 and 11 rows of primordia were excised and

1-5 n

10-
n
0-5-

UJ
vncu

1-5-
os
10 "

0-5- -LP-^L ^
en
NO. MITO:

i i i

1-5 -

1-0 -

0-5 -
w —

5 10 15
DISTANCE (graticule sq.)

Fig. 4. The distribution of mitoticfiguresthrough the condensation and surrounding

dermis after colcemid treatment.
Colcemid: 5-5 h. See text for details. Mitoses were counted in successive columns
of graticule squares at x640 magnification. The mean number of mitoticfiguresper
graticule square (each square 14-7jum wide and 216/^m2 in area) is plotted against
distance from a point 74[j,m lateral to the edge of the condensation. Shaded area:
condensation. Plain area: surrounding dermis.
Numbers of mitoses counted: (A), 223. (B), 236. (C), 261.

IBM B 74
254 D. DAVIDSON
each was trimmed to form a square. A strip of skin, equivalent to between two
and three rows wide, was cut from one side just beyond the most-recently formed
row of primordia and stretched anteroposteriorly. Thus, each pteryla yielded
two pieces of skin: a stretched piece and a complementary, unstretched, control
piece comprising skin medial to the cut and the intact contralateral side. After
between 1 and 2 days in culture, both pieces of skin were measured and the
numbers of primordia per row in the stretched piece and in the control were
compared.
The results are shown in Table 2 and Fig. 5. The principal result is that when
the experiment was done during the development of the first few rows in the
pattern, the number of primordia in the stretched skin was 15 to 45 % greater
than in the control, being roughly proportional of the new length of the skin.
Thus, the region where prospective feather positions could be changed by the

Table 2. The number of primordia formed in stretched as compared with control,

skin
SPECIMEN TREATMENT RESULT
A

Distance of
No. rows at cut from No. of primordia per row
J
H m p f"*"f t"TPtf*h TslCt( iJlr
till It/ \JL LI v l t > U 1 ClV^IA r
operation primordia* Stretched Control
1 0 0-5 46 13 9
2 1 1 28 15 11
3 3 1 34 15 12
4 3 0-5 28 15 13
5 3 0 59 16 12
6 5 0-5 26 15 12
7 5 0-5 41 16 12
8 5 0 46 16 11
9 5 0 50 11 11
10 7 0-5 54 16 12
11 7 0-5 24 13 13
12 7 0 41 16 13
13 7 0 29 11 12
14 9 0-5 44 13(16) 13(13)
15 9 0 38 10 12
16 9 0 23 10 11
17 11 0 30 16 15

* Distance from the most recently formed primordia to the cut beyond which the skin was
stretched: one unit of distance is equivalent to the width of one row.
t Increase in the length of stretched skin as a percentage of length of control skin at the level
of the equivalent row on the contralateral side of the pteryla.
Only the number of primorida in the first row which formed in the stretched skin is shown.
In most cases, however, two rows formed in the stretched skin. In all but one case (no. 14),
approximately the same number of primordia formed in both these rows. For case no. 14 the
numbers of primordia in the second row in the stretched skin and in the equivalent row in the
contralateral control are shown in parentheses.
 Time of determination of feather position in chick 255

oooo^
00 O 0 O
o
Fig. 5. The effect of stretching the skin on the number of primordia per row.
Skin from a single pteryla (case 8, Table 2) was cut immediately lateral to the most
recently formed row (double arrow) just before explanation. The control piece (A)
was explanted without stretching, while the other piece (B) was stretched. Both
pieces are shown after 2 days in culture. (C) and (D) are tracings from photographs
(A) and (B) respectively, and show the positions of well-developed primordia. The
first row formed in (B) comprises 16 primordia (arrow) while the equivalent row in
(A) comprises 11 primordia (arrow).
Magnification of (A-D): X17. Scale bar = 1 mm.

operation moved across the skin immediately ahead of the wave of primordium
morphogenesis.
As more rows formed, the region where the prospective positions of primordia
256 D. DAVIDSON
could be changed by the operation apparently moved further ahead of the visible
wave of primordium formation. This is suggested by two consistent features of
the results. First, the greater the number of rows present at the time of the
operation, the less was the effect of stretching the skin on the number of primor-
dia in the first row in the stretched piece (see Table 2, compare nos. 1, 4, 6, 7,
10 with nos. 11 and 14, and nos. 5, 8, 12 with nos. 9, 13, 15-17). Second, in
specimens with more than about five well-formed rows at the time of the opera-
tion, stretching skin immediately beyond existing primordia had no effect on the
numbers of sites developed in either of the two rows formed in the stretched
piece: in these instances, the distance between primordia was clearly greater than
normal and in a few cases the primordia themselves were abnormally large.

DISCUSSION
Assessment of skin development in culture
The present method of culturing the skin combines several advantages which
are offered individually by other techniques (Hardy, 1951; Schaffer, 1956; Sen-
gel, 1958; Kollar, 1966; McLoughlin, 1961; Dobson, 1967; Novel, 1973; Carinci,
Simonelli, Bubola & Pettazzoni, 1976). In particular, the use of a flat, trans-
parent substratum permits feather development to be monitored continuously
and in detail under the dissecting microscope. Collagen forms a more natural
substratum than those which have been previously employed and is easy to
prepare. Collagen has previously been used to culture isolated epithelia
(Wessells, 1964; Dobson, 1967; Fusenig et al. 1979). The use of liquid nutrient
medium, rather than the incorporation of nutrients into the substratum (Sengel,
1958; Novel, 1973), simplifies the administration of drugs etc., during particular
phases of development.
Although the effects of various tissue extracts on feather development in vitro
have been examined (Sengel, 1958; Novel, 1973), no general assessment of the
development of the feather pattern in cultured skin has been made.
The normal pattern of tissue forces will suffer at least two drastic changes when
the skin is cultured. First, lateral tensions generated by fibroblast traction
(Abercrombie, 1970; Harris, Stopack & Wild, 1981) will be partially relaxed
when the continuity of the skin is broken by excision: evidence of this is provided
by the fact that excised skin shrinks. Second, skin in vivo is part of an expanding
sheet attached to an expanding substratum, whereas cultured skin is isolated
from this dynamic system.
It is likely that many of the abnormalities observed in cultured skin, including
the limited number of rows, the high dermal cell population density, and the
anomolous time course of development (Davidson, 1983), can be understood as
consequences of these two physical effects on the skin. Failure to develop the
normal pattern of forces may also account for the poor organisation of cellular
arrays between condensations in cultured skin since it is possible that these arrays
 Time of determination of feather position in chick 257
result from the alignment of the extracellular matrix by traction fields generated
by cells forming condensations (Weiss, 1959; Harris et al. 1981). The biconvex
shape of primordia during the initial stages of morphogenesis may be maintained
by similar traction forces centrifugal to each condensation. Published work on
other tissues (Phillips & Steinberg, 1969,1978) suggests that without the action
of such forces, the closely packed cells of the condensation would tend to form
a sphere. This may explain why primordia in cultured skin are more rounded and
therefore more elevated than those in vivo. Alternatively, primordium shape
may depend on anchorage of the epidermis to the subdermal mesenchyme by
'anchor filament bundles' (Wessells, 1965; Kischer & Keeter, 1971) which are
scarce in cultured skin. Their scarcity is presumably a result of the destruction
of the base of the dermis, where many of these bundles normally terminate
(Kischer & Keeter, 1971).
One unaccountable finding is the occurrence of mitoses in dermal condensa-
tions in vitro. This contrasts with the virtually complete cessation of DNA
synthesis and mitosis in developing feather and hair primordia in vivo (Wessells,
1965; Wessells & Rossner, 1965). Clearly, therefore, mitotic quiescence is not
a necessary condition for the initial development of feather primordia.

The time of determination of feather position

The principal conclusion to be drawn from the present experiments is that the
visible wave of feather primordium morphogenesis is almost immediately
preceded by a wave of determination at which the positions of feathers are
established row by row. This conclusion follows from the results of the stretching
experiments which show that, during the formation of the first five rows in the
pattern, feather positions are determined close to the time when primordia
appear morphologically. It is assumed that the time when feather positions
become irreversibly established, as defined operationally in these experiments,
indicates the time when they are determined during normal development.
The present results do not establish the precise width of the gap between the
waves of determination and morphogenesis at each point across the pteryla.
They do, however, suggest that the gap is no more than one row wide during the
formation of the first five rows and becomes wider as more rows develop.
During the development of each primordium the epidermal placode forms
before the dermal condensation (Wessells, 1965; Sengel & Rusaouen, 1968).
The data of Sengel and Rusaouen suggests that these two structures form almost
concurrently in the initial row of the spinal pteryla, but, by the time seven rows
have formed, placode development is about two rows ahead of condensation
formation. At first sight, the present results contradict this observation, for
histological analysis did not reveal a difference of more than one row between
epidermal and dermal morphogenesis at any time in the development of the
pteryla. This discrepancy may be due, in part, to the fact that structures with
immature dermal condensations were scored as primordia in the present study.
258 D. DAVIDSON
It may also reflect differences in methodology. Sengel and Rusaouen assayed the
number of rows of placodes in whole epidermal sheets stripped from the dermis.
Conceivably, under these conditions, incipient placodes undergo precocious
morphogenesis by bending of the unattached cell sheet.
In the context of these considerations, the present results display an interesting
correlation with the data of Sengel & Rusaouen (1968). A definitive feather
primordium comprises both a placode and a condensation. Thus, the wave of
morphogenesis of definitive primordia corresponds to the formation of suc-
cessive rows of condensations. Comparing the two sets of data on this basis, the
wave of determination (indicated by the data in Table 2) corresponds closely with
the formation of placodes as revealed by the data of Sengel and Rusaouen. This
interpretation suggests that epidermal feather differentiation begins very close
to the time when the prospective sites of feathers become established.
The present experiments raise two important questions for future investiga-
tion. (1) What is the nature of the organisation in the prospective pteryla which
governs the sequential determination of the feather pattern? (2) Does the
sequential nature of feather pattern development provide a degree of order
which is necessary for the spatial regularity of the final pattern.

It is a pleasure to thank Drs Tom Elsdale and Jonathan Bard for their critical and construc-
tive comments and Ms Allyson Ross for her expert help in preparing the methacrylate his-
tological sections.

REFERENCES
ABERCROMBIE, M. (1946). Estimation of nuclear population from microtome sections. Anat.
Rec. 94, 239-247.
ABERCROMBIE, M. (1970). Contact inhibition in tissue culture. In Vitro. 6, 128-142.
BROTMAN, H. F. (1976). Abnormal morphogenesis of feather structure and pattern in the chick
embryo integument. 1. Macroscopic description. J. exp. Zool. 196, 323-340.
CARINCI, P., SIMONELLI,L.,BUBOLA,G. &PETTAZZONI, G. (1976). An ultrastructural analysis
of chick embryonic skin development in vitro. J. exp. Zool. 197, 389-402.
DAVIDSON, D. R. (1983). The mechanism of feather pattern development in the chick. II.
Control of the sequence of pattern formation. J. Embryol. exp. Morph. 74, 261-273.
DOBSON, J. W. (1967). The differentiation of epidermis. 1. The interrelationship of epidermis
and dermis in embryonic chicken skin. J. Embryol. exp. Morph. 17, 83-105.
EDE, D. (1972). Cell behaviour and embryonic development. Int. J. Neuroscience3,165-174.
ELSDALE, T. R. & BARD, J. B. L. (1972). Collagen substrata for studies on cell behaviour. J.
Cell Biol. 54, 626-637.
FUSENIG, N. E., BREITKREUTZ, D., BOUKAMP, P., LUEDER, M., IRMSCHER, G. & WORST, P. K.
M. (1979). Chemical carcinogenesis in mouse epidermal cell cultures: altered expression of
tissue-specific functions accompanying cell transformation. In Neoplastic Transformation in
Differentiated Epithelial Cells Systems in vitro (eds L. M. Franks & C. B. Wigley). London:
Academic Press.
HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the development of
the chick embryo. J. Morph. 88, 49-93.
HARDY, M. H. (1951). The development of pelage hairs and vibrissae from skin in tissue
culture. Ann. N.Y. Acad. Sci. 53, 546-561.
HARRIS, A. K., STOPACK, D. & WILD, P. (1981). Fibroblast traction as a mechanism for
collagen morphogenesis. Nature 290, 249-251.
 Time of determination of feather position in chick 259
HARRIS, W. F. (1972). Dislocations and control of integumentary patterns in the chick. Devi
Biol. 29, f-l-f-4.
KISCHER, C. W. & KEETER, J. S. (1971). Anchor filament bundles in embryonic skin: origin
and termination. Am. J. Anat. 130, 179-194.
KOLLAR,E. J. (1966). An in vitro study of hair and vibrissae development in embryonic mouse
skin. /. invest. Dermat. 46, 254-262.
LEE, R. L. (1977). 2-Hydroxyethyl methacrylate embedded tissue - a method complementary
to routine paraffin embedding. Med. Lab. Sci. 34, 231-239.
LINSENMAYER, T. (1972a). Control of integumentary patterns in the chick. Devi Biol. 27,
244-271.
LINSENMAYER, T. (19726). Temporal and spatial determination of feather pattern. Devi Biol.
29, M6-M8.
MCLACHLAN, J. C. (1980). The effect of 6-aminonicotinamide on limb development. J. Em-
bryol. exp. Morph. 55, 307-318.
MCLOUGHLIN, C. B. (1961). The importance of mesenchymal factors in the differentiation of
chick epidermis. 1. The differentiation in culture of the isolated epidermis of the embryonic
chick and its response to excess vitamin A. J. Embryol. exp. Morph. 9, 370-384.
NOVEL, G. (1973). Feather pattern stability and reorganization in cultured skin. /. Embryol.
exp. Morph. 30, 605-633.
PHILLIPS, H. M. & STEINBERG, M. S. (1969). Equilibrium measurements of embryonic chick
cell adhesiveness. I. Shape equilibrium in centrifugal fields. Proc. natn. Acad. Sci., U.S.A.
64, 121-127.
PHILLIPS, H. M. & STEINBERG, M. S. (1978). Embryonic tissues as elasticoviscous liquids. I.
Rapid and slow shape changes in centrifuged cell aggregates. J. Cell Sci. 30, 1-20.
SENGEL, P. (1958). Recherches experimentales sur la differenciation des germes plumaires et
du pigment de la peau de l'embryon de poulet en culture in vitro. Ann. des Sci. Nat. Zool.
20, 432-514, plus 4 plates.
SENGEL, P. (1971). The organogenesis and arrangement of cutaneous appendages in birds.
Advances in Morphogenesis. 9, 181-230.
SENGEL, P. (1976). Morphogenesis of Skin. Cambridge, England: Cambridge University
Press.
SENGEL, P. & RUSAOUEN, (1968). Aspects histologiques de la differentiation precoce de
ebauches plumaire chez le poulet. C. r. hebd Seanc. Acad. Sci., serie D. 266, 795-797.
SHAFFER, M. (1956). The culture of organs from the embryonic chick on cellulose-acetate
fabric. Expl Cell Res. 11, 244-248.
STUART, E. S. & MOSCONA, A. A. (1967). Embryonic morphogenesis: role of fibrous lattice
in the development of feathers and feather patterns. Science 157, 947-948.
WEISS, P. (1959). Cellular dynamics. Reviews of Modern Physics 31, 11-20.
WESSELLS, N. K. (1964). Substrate and nutrient effects upon epidermal basal cell orientation
and proliferation. Proc. natn. Acad. Sci., U.S.A. 52, 252-259.
WESSELLS, N. K. (1965). Morphology and proliferation during early feather development.
Devi Biol. 12, 131-153.
WESSELLS, N. K. & ROSSNER, K. D. (1965). Nonproliferation in dermal condensations of
mouse vibrissae and pelage hairs. Devi Biol. 12, 419-433.
ZEEMAN, E. C. (1974). Primary and secondary waves in developmental biology. Lectures in
Mathematics in the Life Sciences, Volume 4, pp. 69-161. Rhode Island: Mathematical
society.

(Received 24 November 1982)