/ . Embryol. exp. Morph. Vol. 53, pp. 245-267, 1979
Printed in Great Britain © Company of Biologists Limited 1977
245
Somitogenesis in amphibia
II. Origins in early embryogenesis of two factors
involved in somite specification
By TOM ELSDALE 1 AND MURRAY PEARSON 2
From the Medical Research Council, Clinical and Population Cytogenetics Unit
Western General Hospital, Edinburgh
SUMMARY
A somite pre-pattern is established shortly before visible segmentation. The pre-pattern
results from the interaction of two components: a wave of cell behavioural change that passes
along the axis, and, an underlying co-ordination of the cells that is the basis for their association into large somite-sized groupings. The evidence is derived from studies of the zones of
abnormal segmentation that follow temperature shocks delivered between the neurula and
tail-bud stages (Pearson & Elsdale, 1979).
Temperature shock given earlier at the mid-gastrula stage is however ineffective in inducing
abnormalities in somitogenesis. Shocks given before the mid-gastrula stage reveal a prior
period of sensitivity stretching back into the blastula. Thus early and late sensitive periods can
be defined separated by a short refactory period. Quite different patterns in the distribution of
somite abnormalities characterize the results of shock during the two sensitive periods,
suggesting different aetiologies.
It is concluded that the wave of rapid cell change is set up early in embryogenesis during the
blastula stage, and each cell of the prospective paraxial mesoderm carries a determination to
change after a specific length of time, i.e. a countdown is set in each cell. As a result of the
movements of gastrulation, the prospective paraxial mesoderm cells become laid out along
the axis of the neurula in the order (antero-posterior sequence) in which they will change. The
achievement of the correct redistribution of the cells depends crucially on the conservation of
the sequence in the blastula by the maintenance of topological integrity throughout gastrulation. It is suggested that early shock disturbs gastrulation movements, causing some mixing up
of the cells resulting in incoherence of the wavefront.
Whereas early shocks are thus assumed to affect the wave, the evidence suggests that late
shock undoes co-ordination. It is concluded therefore that co-ordination is established later,
after the refractory period, around the late gastrula stage.
INTRODUCTION
Somitogenesis is governed by two time intervals: the long time taken for the
somites to form from first to last is partitioned into equal short intervals between
the formation of one somite and the next.
1
Author's address: MRC Clinical and population Cytogenetics Unit, Western General
Hospital, Crewe Road, Edinburgh, U.K.
2
Author's address: 30 Albion Hill, Brighton, Sussex, U.K.
246
T ELSDALE AND M. PEARSON
From experiments employing temperature shocks delivered to embryos about
to form and forming their somites, and the analysis of the location, extent and
anatomy of the characteristic abnormal zones developed in the somite files
following this treatment, the following scheme has emerged. The longer interval
is defined by the time taken to traverse the axis by a wave of rapid cell change
moving in advance of somite segmentation. A second component, a co-ordination of the cells, contributes to the definition of the shorter intervals; a momentary coupling of these two components as the wavefront moves along the axis,
establishes a somite pre-pattern which manifests as visible segmentation after a
fixed time delay (Pearson & Elsdale, 1979).
This scheme has basic similarities with a theoretical model of somitogenesis
which shows how a wave of change passing through a field of co-ordinated cells
can function like an escapement mechanism in a clock to provide for a repetitive
series of discontinuous events (Cooke & Zeeman, 1976).
In order to shed light on the origins in early embryogenesis of the two components that interact to specify the segmental pattern, we have studied embryos
temperature-shocked during gastrulation.
Three experiments have been performed. First, a series of shocks to different
gastrula stages shows that somite abnormalities are induced after shock to the
younger gastrula but not the older. Furthermore, the distribution of abnormal
somites in the former is different from what is observed in embryos shocked at
the neurula and post-neurula stages. Second, the refractory period to temperature shock during later gastrulation indicated by the previous experiment, is
investigated with longer duration shocks. We show in our first paper of this
series how shocks to neurula and post-neurula stages induce somite abnormalities and in addition an enduring hidden effect revealed by a second shock; in
a third experiment it is shown that a shock during the refractory period having
no visible effect, has no hidden effect either. It is argued that a genuine refractory
period separating earlier and later periods of sensitivity to temperature shock
indicates disturbances to different components in segmental specification.
MATERIALS AND METHODS
Rana temporaria embryos collected in season from natural habitats have three
advantages over Xenopus: (1) each spawning provides a large population of
synchronously developing embryos, (2) virtually all the embryos develop normally and (3) Rana embryos develop more slowly than Xenopus. The fact that
Rana embryos are nearly twice the size of Xenopus is an added convenience.
Clutches of early cleavage stage Rana were collected in Perthshire and kept
at 6 °C in a cooled incubator in water collected with embryos. The day before
each experiment embryos were brought to room temperature to allow a period
of adaptation to the higher temperature of 19 ± 1 °C. During most experiments the
embryos were maintained in pond water, some use was made of tap water that
Somitogenesis in amphibia. II
SulhstagcV
(SS)
1
Substage I = Shumway
Substage 10 = Sluimway
Substage 15 = Sliumway
Substage 19 = Slunnway
247
J
st.
st.
st.
st.
10
11
12
13
Fig. 1. The course of gastrulation in Rana temporaria reared at 19 °C. Each sketch
shows the progress made by the blastopore viewed externally at one of the substages at which a 9-min temperature shock was delivered. A selection of these substages is illustrated. How these substages accord with Shumway's (1940) staging of
gastrulation in Rana pipiens is indicated.
had stood for a week under continous aeration. Embryos denuded of their jelly
were re-examined immediately before shock in order that visibly damaged
embryos could be discarded. To give a heat shock, embryos were transferred
with minimal carry over of water to glass bottles containing 125 ml of pond
water standing in a water bath at 37 + 0-2 °C. After the desired time at 37 °C,
embryos were transferred to a larger body of water at room temperature, and
reared until ready for scoring. In this way a standard 9-min duration shock was
repeatable within an accuracy of 10 sec (2%).
Somite files were revealed for scoring by stripping the skin from the embryos.
Embryos were transferred to a solution containing 0-5 % potassium dichromate
and 2-5 % glacial acetic acid (Smith's fixative without formaldehyde). After a
few minutes in this solution the skin becomes strippable and remains so for some
10-15 min before turning brittle. Files were scored and photographed under a
stereo microscope using incident oblique illumination.
Scanning electron microscopy and histological preparation employed the
same techniques used by Pearson & Elsdale (1979).
At room temperature 19 ± 1 °C it takes 22 h for Rana embryos to develop
from the initial gastrula to the neural plate stage. This period is conveniently
divided into 22 substages marking the progress of gastrulation hour by hour.
The external appearance of the blastopore at representative substages is illustrated in Fig. 1 which gives a synoptic view of the time course of gastrulation.
Three of these substages are reference substages, defined independently of the series
on external criteria alone, these are: SS 1 the initial gastrula, SS 12 coincident
with the completion of the circular blastopore, and SS 18 at which the yolk
248
T ELSDALE AND M. PEARSON
button disappears from view, a convenient marker for the end of gastrulation. Fig. 1
shows that the lateral extension of the blastopore to form the blastoporal circle
is a relatively slow process, whereas the diminution of the yolk plug proceeds
rapidly thereafter. Following common usage, SS 12 is referred to as the midgastrula stage although it is actually the stage two thirds of the time through
gastrulation;
EXPERIMENTAL PROCEDURES AND RESULTS
1. 9-min temperature shocks during gastrulation
Embryos developing synchronously at room temperature were divided into
lots, each receiving a shock at one of the substages of gastrulation. The experiment comprised two series of shocks, one ovulation was used throughout the
first series, and another throughout the second (Table 1).
Series 1
The first series comprised a control, and 16 lots receiving temperature shock.
Fifteen of the later provided a series of shocks from SS 1 to SS 15. A late shock
at SS 22 provided an overlapping comparison with previous work. Embryos were
scored for abnormal gastrulation and abnormal somitogenesis. When possible
each somite was scored in both files of the same embryo at around the 25-somite
stage; these scores are the basis for Fig. 3.
For descriptive purposes the series is divided into four subseries. After noting
the salient features in each, the subseries are described in detail.
Subseries 1: SS 1-SS 6. Gastrulation and somitogenesis grossly abnormal.
Subseries 2: SS 7-SS 11. Gastrulation and somitogenesis improving. Posterior
relegation of abnormal somites.
Subseries 3: SS 12-SS 15. Normal gastrulation. Trivial somite irregularities.
Subseries 4: SS 22. Zone of abnormal segmentation in all embryos extending
over first several somites. Characteristic effect of shock to neurula stage.
Shocks to SS 1-SS 6
Only in this subseries did a significant number of embryos die without developing a primary axis, and most of this mortality was associated with shocks to the
first three substages. This result already points to the trend characterizing the
series as a whole, namely, the later the 9-min shock is given, the less the effect.
The overall result, although variable, is a severe disturbance to both gastrulation and somitogenesis, without however a close correlation between the severity
of these two effects within the same embryo. Although most surviving embryos
were stunted, a minority showed near normal axial extension although their
somite files were grossly abnormal.
(a) Gastrulation: The subseries shows that the earlier a temperature shock is
given during gastrulation the greater the likelihood that the external course of
19
8
—
—
12
7
—
—
—
0
24
17
—
—
20
2
6
—
18
—
—
42
—
0
0
26
21
—
—
21
0
0
Series 1
7
22
0
9
6
28
0
14
0
0
24
2
0
24
13
0
0
11
12
0
0
10
42
0
21
0
0
12
20
4
12
0
0
13
28
0
14
0
0
14
26
No. of embros shocked
No. dead at time of
scoring
No. of files scored somite
by somite
Lots (Substage)
JO
1
18
20
2
(15)
10
0
1
(14)
18
2
11 -
Series 2
3
(16)
18
10
1
4
(17)
5
12
0
24
22
6
(19)
0
n
08)
0
13
0
0
15
All surviving embryos in Lots 1 and 2 were grossly abnormal and the somite files unscorable. Lots 4, 5 and 8 were not scored.
No. of embryos
No. dead gastrulae
No. survived gastrulation
dead at time of scoring
Files damaged during
stripping, not scored
No. of files scored
somite by somite
Lots substage shocked
Table 1. Scheme of temperature shocks to gastrula stages
49
16
8
0
1
25
0
0
Control
22
0
11
0
0
22 Control
250
T. ELSDALE AND M. PEARSON
Fig. 2. Each of these stripped embryos from Series I received a 9-min temperature
shock early in gastrulation. (a) Shock delivered to SS 1. Inhibition of segmentation.
Somites are not distinguished in this grossly abnormal specimen. Some faintly
demarcated furrows are alone visible, (b) Shock delivered to SS 3. Abnormal segmentation. There is a partial inhibition of segmentation anteriorly, chaotic segmentation (Ch) posteriorly. Frayed ventral edges (/) indicate indistinct demarcation
of the somite mesoderm from the lateral plate, (c) Shock delivered to SS 6. Posterior
restriction of abnormal somites is evident in this embryo. The anterior somites are
spared, behind them somitogenesis is abnormal.
Somitogenesis in amphibia. II
251
gastrulation is arrested at the circular blastopore stage. The great majority of
embryos so arrested however subsequently developed neural plates in the presence of a large persistent yolk plug and various degrees of spina bifida.
Persistence of the yolk plug was correlated with changes in the superficial
pigmented ectoderm observed within a day of shock, a pronounced crinkling
and folding of the ectoderm was observed that remained a feature of this tissue
after its translocation to the ventro-lateral surface of the later embryo.
(b) Somitogenesis: Somitogenesis in this subseries was so grossly abnormal
that it was often impossible to score files somite by somite. In lots 1 and 2, a
great many files were abnormal throughout. For comparison with later subseries, scores are given for SS 3 and SS 6 only (Fig. 3).
The appearance of abnormalities can be described under the following heads,
with the qualification that the appearances described vary widely in severity
between affected embryos.
(1) Inhibition. A total inhibition of segmentation was invariably associated
with severe stunting and an arrest of development after neurulation resulting in
disintegration. Total inhibition was however rarely encountered. Much more
common was a premature cessation of somitogenesis after the formation of a
variable number of abnormal somites (Fig. 2a).
(2) Faint somite boundaries, suggesting attenuation of the segmental impulse,
was commonly seen. There was often a gradient along the axis, with posterior
somites less clearly delineated than anterior.
(3) Local inhibition. Local failures of segmentation, bounded by regions of
more normal segmentation, were common. Such local areas were generally
followed by gross abnormalities.
(4) Chaotic segmentation. The impression here is of fragmentation and disorganization of the segmenting mesodermal units (Fig. 2 b). The excess fissures
between the small irregularly segmented groups of cells often ran in all directions.
(5) Frayed edges, a frequent concomitant of other abnormalities, where the
ventral border of somitic mesoderm was imperfectly demarcated from the lateral
plate (Fig. 2b).
(6) 'Mistakes'. Trivial inconsistencies, to be described later.
Shocks to SS7-SS 11
There is no discontinuity between this subseries and the first. Most of the
embryos gastrulated normally and the pigmented ectoderm remained smooth.
The somite scores in Fig. 3 show the abnormalities in retreat; not only are there
fewer abnormal somites, there is also a clear trend towards the sparing of anterior somites and a confinement of abnormalities to more posterior levels,
(Fig. 2 c). A further trend not brought out in Fig. 3 is towards a decline in the
severity of abnormalities.
252
T. ELSDALE AND M. PEARSON
SS3
100-1
SS 12
I I I 11 I Ml I I I I I 11 11 I I 11 11 I I
Control
SS 11
50-
5025T T T I T 1 1 I I I I 11 I I I I I I I I I I H
1
5
10
15
20
I 1 1 1 11 I I I I I I I 11 I I 1 1 I I 1
25
1 5
Somite number
10
15
20
25
Fig. 3.9-min temperature shocks during gastrulation, Series 1. Results. Embryos were
stripped and scored between the 20 and 25 somite stages. Each somite in each embryo
was numbered 1, 2, 3...according to its location in the sequence of somites, and
scored as either normal or abnormal. The degree of abnormality was not taken into
account. These scores provide the primary data. Reference to Table 1 shows that a
significant number of embryos shocked at SS 3 and SS 6 died before their batches
were scored. The data from these substages are therefore biased. The remainder of the
data is not biased. For each numbered somite, the scores, from all the embryos
shocked at the same substage, were summed. The proportion of somites abnormal
within each of these summations was converted to a percentage. These percentages
are displayed in histograms. The trend towards fewer abnormalities and the sparing
of anterior somites on passing from SS 6 to SS 11 is clear. From SS 12 to SS 15
abnormal scores are no higher than controls, and like the control scores reflect merely
trivial irregularities of the type termed 'mistakes', see page 258 and Fig. 7. The
figure shows in striking manner the sensitivity of the earlier gastrula to disturbance
by temperature shock, and how as development proceeds the gastrula becomes more
and more resistant, and finally refractory around SS 12-15. By SS 22 however the
embryo is again easily disturbed, but now the distribution of abnormalities is quite
different from before, and they are concentrated into an anterior abnormal zone.
Somitogenesis in amphibia. II
253
Shocks to SS J2-SS 15
The subseries is characterized by normal somite files; such anomalies as were
observed were trivial irregularities classed as mistakes, to be described later.
This subseries demonstrates the refractory period to temperature shock.
Shock to SS 22
The effect of a shock to embryos about to form and forming their somites is
much more predictable than the effect of a shock to earlier stages; abnormal
segmentation is confined to a limited zone the length of which is proportional to
the duration of shock (Elsdale, Pearson & Whitehead, 1976). The location of the
abnormal zone is precisely correlated with the progress of segmentation; a shock
delivered shortly before somitogenesis commences, as in the present case, induces
an abnormal zone including the very first somites; following a shock to embryos
in the course of forming their somites, the zone invariably commences three to
four somites posterior to the last somite formed at the time of shock. The abnormal zone is characterized by chaotic segmentation and an excess of fissures
anteriorly, and a gradual amelioration and return to normal segmentation
posteriorly.
1B. Series 2
The first series gives a detailed picture of the transition from an early gastrula extremely sensitive to disturbance by temperature shock to a 'mid-gastrula'
that is refractory. The gap between SS 15 and SS 22 however leaves undocumented the transition from the refractory stage to the sensitive early neurula.
The series comprised a control and six lots shocked at hourly intervals from
SS 14 to SS 19. The results are presented in Fig. 4 and Table 2. Abnormalities
among the first formed somites are significant in embryos shocked at SS 16-17,
and typical abnormal zones are present in embryos shocked at SS 19.
1C. Summary of gastrula shocks
A refractory period within the latter half of gastrulation during which temperature shock does not induce somite abnormalities is flanked by preceding
and succeeding periods of greater reactivity.
Shocks delivered at the time the blastopore is initiated induce abnormal
gastrulation and persistence of the yolk plug, and our data show that the first
25 somites are about equally at risk to severe disturbance. The trend exhibited in
embryos shocked closer to the mid-gastrula stage is towards normal gastrulation,
a progressive sparing of anterior somites and a confinement of abnormalities to
more posterior levels concomitant with a declining incidence and severity. This
trend culminates in the absence of abnormalities from embryos shocked at the
beginning of the refractory period.
The trend exhibited, following a succession of shocks delivered as the late
17
EMB53
254
T. ELSDALE AND M. PEARSON
lOO-i
Control
100-
SS 17
50 -
50-
I i i i i 1 1 I [ 1 1 1 1 r 1 1 1 i 11
100-1
SS 18
SS 14
50-
I 11 I 11 I I I I I | i 1 1 I I
100 -i
SS 19
SS 15
50-
T T T I T T T I 1111 ii i II
100-1
SS 16
100 1
SS22
50 "
50-
TTTTTTT 11 II 1 I 11
5
10
15
10
15
20
Somite number
Fig. 4. 9-min temperature shocks during gastrulation, Series 2. Results, see in conjunction with Table 2. The format is the same as Fig. 3. Embryos were stripped and
scored between the 10- and 15-somite stages, and for this reason the data are not
strictly comparable with the data from thefirstseries. By the 20- to 25-somite stages at
which the first series embryos were scored the first three somites are becoming
partially obscured by developments in the ear region. As shocks around the conclusion of the refractory period were expected to induce abnormalities in these first
formed somites, the second series was scored at an earlier stage at which these somites
could be viewed without dissection after stripping. The drawback to early scoring is
that the segmental boundaries have not completely settled down and their appearance
often presents small irregularities absent from older embryos. For this reason we have
a rather high control count amounting to about one abnormal somite per file. The
first series results revealed a refractory period to the standard temperature shock
commencing around SS 11 and continuing to at least SS 15. The second series results
show that in fact an SS 15 embryo is already approaching the end of the refractory
period, and a shock at SS 17 reliably induces abnormalities among the first two or
three somites. As embryos develop towards the neural plate stage, their susceptibility
to shock further increases, and the abnormal zones induced by the 9-min shock
lengthen.
Somitogenesis in amphibia. II
255
Table 2. Results of second series of shocks to gastrula substage also including the
data from Series 1, substage 22
Substage
No. offilesscored
No. of files without
abnormalities
Mean length of abnormal
zone: mean somites
counted from somite 1
before first normal
somite
14
15
16
Series 1
17
18
19
22
Control,
20
10
18
3
18
0
18
0
24
0
22
0
16
5
0-6
(•astmhir abnormalities
21
2-5
22
0
3-9
50
6-9
5-5
0-8
Ncurula type abnormalities
5-0-1
h6 e
o
4-0 -
£
.
d
•4
E
3 0 -
o
20 -
~ 1
r
7
I
8
I I I I I 1
I I 1 I T T T I
9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 192 0 2 1 2 2
Substages o f development
Fig. 5. The refractory period. Data from both series are employed to present the refractory period as a basin with sloping sides. Solid circles, Series 1: For each substage
the total abnormal somites scored has been divided by the number offilesscored, to
give a figure for the mean number of abnormal somites per file. The appropriate
axis is on the left. The base line represents the control for the series. Open circles,
Series 2: The data along the bottom row of Table 2 are used. The appropriately
scaled axis is on the right. The dashed horizontal line represents the control for the
series. The figure shows that the refractory period is centred around SS 13.
17-2
256
T. ELSDALE AND M. PEARSON
(b)
Fig. 6. An embryo stripped at the 35-somite stage following a 20-min temperature
shock during the refractory period (SS 13). (a) Anterior trunk; (b) part of tail. These
pictures could pass for illustrations of a normal untreated embryo.
gastrula develops towards the neural plate stage, is towards increasing abnormalities among thefirstformed somites culminating in the typical abnormal zones
characteristic of later shocks.
The two periods of reactivity to temperature shock separated by the refractory period are distinguished by quite different patterns in the distribution of
segmental anomalies.
For practical purposes the refractory period is preferably delimited in terms
Somitogenesis in amphibia. II
257
Table 3. The frequency of mistakes counted at the 35-somite stage following temperature shocks delivered at SS 13 during the refractory period
Duration of shock
(min)
10
15
20
No. of
scored
28
28
24
files
Mistakes
Mean mistakes
per file
12
48
185
0-4
1-7
7-7
of our reference substages based on unequivocal external criteria. It needs to be
kept in mind however that embryos neither enter nor leave the refractory period
abruptly; in Fig. 5 the data are used to present the refractory period as a basin with
sloping sides. Reference substage 12 the 'mid-gastrula' stage is a satisfactory
marker for the beginning of the RP, reference substage 18 however is too advanced to represent the end of the period which corresponds with SS 16. In
Rana embryos reared at 19 °C the refractory period can be said to last about 4 h.
2. 20-min shock during the refractory period (RP)
Shocks of long duration were employed in order to discover whether embryos
passing through the RP were genuinely incapable of reacting to shock by abnormal segmentation, or whether they were merely more resistant to shock.
Early gastrulae are killed by a 20-min shock. Delivered after the RP a shock
of this duration is not lethal, but induces very long abnormal zones; indeed only
when the abnormal zone commences within the first 20 somites is there an
eventual return to normal segmentation caudally.
Delivered at SS 13 (RP), around a quarter of the embryos are killed by a
20-min shock, showing it is near the maximum that can be used in a meaningful
experiment employing embryos at this stage. The immediate effect of temperature shock is an arrest of development the duration of which is proportional to
the length of the shock. Subsequently, embryos resume development at the
normal rate. Survivors of a 20-min shock at SS 13 resume development after a
protracted arrest lasting nearly three days; they nevertheless continue to grow
normally forming an extended body axis and tail, apart from a kink in the tail
that possibly indicates a local effect on notochordal development. Segmentation
appears normal in these embryos and is nowhere seriously disturbed (Fig. 6).
The sole irregularities are those classed as mistakes. These latter are increased to
an average of 7-7 per file, a 40-fold increase over controls (Table 3).
Mistakes (Fig. 7) refer to trivial irregularities reflecting nothing more than a
failure in the correct registration of the dorsal and ventral half somites. The
commonest appearances are a Y formation when two adjacent dorsal half somites are confluent with a single ventral half somite, and the reverse situation
258
T. ELSDALE AND M. PEARSON
Fig. 7. Mistakes. Short lengths from the tails of two embryos stripped between the
25- and 30-somite stages. Left is anterior, top is dorsal, (a) The third complete segment counting from the left is a Y formation, two dorsal half somites are fused to
one ventral half somite. This is counted as one mistake. Adjacent to the right is an
inverted Y formation. The two mistakes compensate, for together they comprise
three dorsal half somites and three ventral, and the pattern following is not put out.
(b) Two similar mistakes to those in (a) but here separated by one whole somite. The
imbalance due to the extra dorsal half somite created by the first mistake persists
through an extra somite before being corrected by the second mistake. Mistakes do
not always occur in pairs like this.
giving an inverted Y. Occasionally a floating half somite may occur; three half
somites joined to one has been observed but is unusual.
In the case of late shocks (delivered after the RP) inducing a limited abnormal
zone, the length of this zone is proportional to the duration of shock; in response
to minimal shock the abnormal zone reduces to a mistake. However, mistakes
occur in other situations where they cannot be considered as reduced abnormalities. Mistakes are a natural occurrence, routinely observed both in controls
reared in the laboratory and in embryos reared in their natural habitats. Among
the latter, around one file in six bears one or more mistakes and it is interesting
that these are not observed in the first 13 somites that alone persist after metamorphosis. Thus mistakes can hardly be considered abnormal in themselves.
The incidence of mistakes can be increased by procedures other than heat
shock that do not induce somite abnormalities. Mistakes are always increased
following temperature shock; thus, for example, there is an increased incidence
of mistakes posterior to the abnormal zones, within the remainder of the files
Somitogenesis in amphibia. II
259
Table 4. Double temperature shock experiment
Stage at shock
^^i l mnpr
1™ 1411IUWL
Experiment
of files
scored
First
Abnormal
abnormal
somites,
somite of
second abnorsecond abnor- mal zone in
experimentals
mal zone
A
A
f
N 1
First
Second
Mean
no.
S.D.
Mean
no.
S.D.
la
Experimental
Control
16
Experimental
Control
2a
Experimental
Control
26
Experimental
Control
34
27
1st som.
9th som.
9th som.
None
13-3
12-0
0-9
0-8
2-1
5-5
1-5
1-9
41
30
SS 15 RP
9th som.
9th som.
None
12-9
11-3
0-9
0-7
4-7
50
2-2
1-6
15
14
1 st som.
10 som.
10 som.
None
13-9
13-9
0-6
1-1
4-4
6-9
1-1
11
26
22
SS 15 RP
4 som.
4 som.
None
60
4-9
1-6
0-8
6-4
5-8
1-7
1-7
In experiment 1 second shocks were delivered to comparable stages as explained in the text,
and in experiment 2 the interval between the first and second shocks was the same in both
halves of the experiment. The (a) half of each experiment comprises embryos receiving the
first of two shocks at the 1-somite stage; the (b) half comprises embryos receiving the first of
their two shocks during the RP. At least two normal somites intervened between the end of the
first abnormal zone and the beginning of the second along the same file.
otherwise segmenting normally following a late shock. In neither of these two
cases can extra mistakes be considered reduced abnormalities.
It appears that mistakes reflect an aspect of somitogenesis that has remained
unstable under the influence of natural selection, and is therefore easily disturbed. For this reason we do not consider extra mistakes in the absence of
abnormalities as an effect of a severe temperature shock continuous with the
production of gross abnormalities.
The significance of the experiment is that it clearly demonstrates the extraordinary refractoriness on the part of embryos at the mid-gastrula stage to the
induction of somite abnormalities, and thereby incidentally contrives an unobscured window onto the raised background of mistakes that invariably
accompanies temperature shock.
3. Two temperature shocks to the same embryo
By giving two temperature shocks to the same embryo, we have obtained two
abnormal zones along the somite files separated by normal somites (Pearson &
Elsdale, 1979). The interesting fact emerged from these experiments that the
second abnormal zone was invariably only half as long as expected, that is to say
260
T. ELSDALE AND M. PEARSON
Experiment 1
First shock
First shock
Second shock
Second shock
0)
50-
• i I i i I I I I i i I i i I
0 1 2 3 4 5 6 7 8 910111213
Experiment 2
1 1 1 1 1 i
0 1 2 3 4 5 6 7 8 910111213
Number of somites affected
0 1 2 3 4 5 6 7 8 910111213
I
T } 1
T
i
l I 1 I I
0 1 2 3 4 5 6 7 8 910111213
"•Number of somites affected
Fig. 8. Double temperature shock experiment. The primary data consist of estimates
of the number of somites within abnormal zones induced by 8-min temperature
shocks. Estimates were made according to Pearson & Elsdale (1979). In the case of
embryos with two abnormal zones as a result of two shocks, the first zone was
ignored as only the length of the second was relevant to the experiment. A few
embryos were discarded in which the two somites, immediately preceding the second
abnormal zone, were not unequivocally normal. In order to compare the lengths of
the abnormal zones in two classes of embryos it is necessary to take into account
the variation within each class. The figure therefore displays distributions showing
the proportion of files in each class having abnormal zones of length 0,1,2,.. .somites.
The proportions are expressed as percentages. Each graph compares two such distributions. The distribution bounded by the solid line, in each case refers to abnormal
zones induced by a second shock. The accompanying distribution bounded by the
dashed line refers to the corresponding control embryos that did not receive the earlier
of the two shocks. Of significance is the extent to which the two distributions overlap;
the overlaps are therefore cross hatched. The protective effect of a temperature shock
is demonstrated by a leftward shift of the distribution bounded by the solid line in
relation to the distribution bounded by the dashed line. Such an effect is demonstrated in the two graphs on the left (la and 2a) referring to embryos receiving a
first shock at the 1-somite stage. In contrast the distributions on the right (1 b and 26)
overlap over most of their ranges indicating that a shock during the refractory period
has no protective effect.
Somitogenesis in amphibia. II
261
half as long as the zones in control embryos that had not received the earlier of
the two shocks. This result demonstrated an enduring effect of temperature
shock persisting after segmentation had returned to normal, conferring a partial
protection against a subsequent shock. Does a shock during the refractory period
confer a similar protection?
We compared the repercussions of a shock delivered during the RP with the
repercussions of a shock delivered at the first somite stage. Two experiments
were performed. In the first, the test shocks were given to both groups at the
8-somite stage and the interval between the first and second shocks was different
for the two groups. In the second experiment the interval between the two shocks
was the same for both groups and the embryos received their second shocks at
different stages. The results are presented in Table 4 and Fig. 8. In contrast to a
shock at the first somite stage, a shock during the RP has no influence on the
effect of a subsequent shock.
It is concluded therefore that a shock during the RP neither induces abnormalities along the somite file, nor has the enduring protective effect of a shock
delivered to later stages.
DISCUSSION
The first paper in this series (Pearson & Elsdale, 1979) dealt in detail with the
effect of a late temperature shock (delivered after the RP). Following a section
on the significance of the RP, the bulk of this discussion is devoted to the effect
of early temperature shock (delivered before the RP).
The refractory period
We have suggested on the basis of previous experiment a mechanism by which
late shocks cause abnormalities in segmentation. It was assumed that late shocks
disturb all of the paraxial mesoderm not yet committed to segmentation, by
undoing a co-ordination of the cells essential for normal patterning. The disturbed tissue slowly recovers. Crucial is the time available for recovery before the
arrival of the wave of rapid cell change. The somite pre-pattern established
with the passing of the wave reflects the state of co-ordination of the cells, and
disturbed co-ordination at this time will leave its indelible record in the somite
file (Fig. 15, Pearson & Elsdale, 1979).
On the basis of this scheme it might be supposed that shocks during the RP
produced no somite abnormalities because recovery is invariably complete
before the pre-pattern of the first somites is established. There is no time for
such recovery. The first somite forms as the neural folds close; utilizing the data
displayed in fig. 7 of Pearson & Elsdale (1979), it can be estimated that the
determination of the first somite follows within a few hours of the RP.
An alternative possibility is that temperature shock during the RP is not
registered by the paraxial mesoderm. Two independent tests can be applied.
First, one can look for somite abnormalities induced by shock. Second, the
262
T. ELSDALE AND M. PEARSON
enduring protective effect of a shock can be investigated by a double shock experiment. Negative results from both tests indicate the unresponsiveness of the
prospective paraxial tissue to a temperature shock during the RP.
Having previously concluded (Pearson & Elsdale, 1979) that late shock disturbs the co-ordination of the cells essential for normal patterning, while leaving
the wave untouched, we infer that co-ordination is not established until the
end of the RP. It follows that co-ordination cannot be the target of early shock;
this target must be absent from the beginning of the RP. Moreover, separate
aetiologies are strongly suggested by the characteristically different distributions
of abnormalities along the somite files associated with early and late shocks.
The RP is interpreted therefore as an interval between two periods of differing
reaction to temperature shock, in which the occasion for the earlier has passed
and the conditions for the later have yet to mature.
The effect of early temperature shock
1. No recovery after early shock
We next enquire whether the data from early shock provide the same sort of
evidence for a recovery process as do the data from late shocks.
A recovery process was postulated to account for the invariable finding that
the longer the interval between a late shock and the formation of the somite prepattern coincident with the advance of the wave of rapid cell change, the less
severe the induced abnormalities in segmentation, exemplified by the gradual
return to normal segmentation behind the anterior of the abnormal zone
(Elsdale et al. 1976). We enquire therefore whether a similar rule characterizes
the result of early shock.
The data show that the shock of the same duration induces more numerous
and more severe somite abnormalities delivered to the early gastrula than delivered to progressively later gastrula stages, although the time available for
recovery is longest in the case of the former. Further, through SS 1 to SS 12 we
observe a progressive sparing of anterior somites and a relegation of abnormalities to the more posterior of the first 20-25 somites, although anterior cells have
less time to recover than posterior cells. These results do not merely offer no
support for a recovery process, they provide a compelling contra-indication and
justify the rather categorical conclusion that there is no recovery after early
shock. Using minimal shocks it is likely that Cooke obtained a less convincing
contra-indication than we did using more severe shocks, and he is not deterred on
the basis of his results on Xenopus from envisaging a slow recovery after early
shock (Cooke, 1978).
2. Delimiting the target of early shock
Unpublished observations reveal that the blastula is even more sensitive to
heat shock than the early gastrula. Observations over a 16 h period show that the
sensitivity of the late blastula declines steadily during the approach to gastru-
Somitogenesis in amphibia. II
263
lation. Seen from this perspective our results indicate that as gastrulation gets
under way sensitivity to heat shock dies away sharply. In our Series 1 we are
observing the tail end of a sensitive period not yet adequately explored, extending back into the blastula.
The target in the earliest gastrula comprises the precursors of no more than
the first 30 somites (Cooke, 1978). As gastrulation proceeds the target area
(in terms of prospective somites) contracts as the precursors of the anterior
somites fall out of the target. We can picture a change from target to non-target
overtaking the prospective somite mesoderm in the direction of the future
anterior posterior axis. At the same time as the target is thus contracting, the
number and severity of the abnormalities inducible within the residual target
area is declining. We can picture therefore a concomitant attenuation or dilution of targets. By the mid-gastrula stage marking entry into the R.P., the target
it seems has been eliminated as a result of the twin processes of contraction and
attenuation.
We next seek to locate this target in the early embryo. In the absence of detailed knowledge of gastrulation in Rana we are forced to make assumptions on
the basis of knowledge of other anurans, expecially Xenopus. A tentative correlation of gastrulation movements with the contraction and attenuation of the
target of early shock and the timing of the R.P. suggests that the prospective
somite mesoderm is especially sensitive prior to involution, and perhaps still
somewhat sensitive during involution, whereas the tissue joining the mesodermal
mantle following involution is no longer sensitive. However, the decline in
sensitivity to temperature shock during the first two-thirds of gastrulation is
much too rapid to correlate with the course of involution of the prospective somite
tissue. It is Raymond Keller's educated guess (personal communication) that
perhaps no more than the first five prospective somites have involuted by the
mid-gastrula stage marking the beginning of R.P. Even allowing considerable
latitude to Keller's estimate, we must envisage a sharp decline in sensitivity to
temperature shock of the prospective somite tissue prior to its involution, a
decline that may indeed already be under way in the late blastula.
These considerations raise the question whether it is the cells themselves that
are disturbed by heat shock independently of the movements of gastrulation, or
whether it is these movements that are disturbed in which case it is the consequent irregularities in the spatial distribution of the cells that are eventually
translated into abnormalities in segmentation.
3. Abnormal gastrulation after early shock
Although the direct effect of heat shock on the prospective somite mesoderm
is not observable externally, observation of the striking and characteristic effect
of early shock on the superficial tissues of the embryo is of interest because it
indicates that the same early shocks that lead to a disturbance of somitogenesis
also interfere with the morphogenetic movements of gastrulation.
264
T. ELSDALE AND M. PEARSON
Epiboly, whereby the yolk is enclosed by a ventral streaming of the dorsal
ectoderm, occurs during the second half of gastrulation following the formation
of the circular blastopore. The inhibition of epiboly after shock to the early
gastrula leading to persistent yolk plug and spina bifida has been described. It
appears that early shock induces a premature and unco-ordinated expansion of
the dorsal ectoderm manifested by folding and crinkling of the surface, with the
further result that the co-ordinated ventral streaming movements by which the
yolk is enclosed are incapacitated. It is interesting that the gastrulation advances
towards the R.P. as does the somite-forming system.
The disturbance of ectodermal epiboly following early shock may point to a
general disturbance of the gastrulation movements. However, somite abnormalities may be the only abnormal indications in embryos otherwise comparable
to controls after allowing for the retardation which invariably follows early
shock. In such embryos there can have been no gross disturbance of gastrulation;
at most, local irregularities in the redistribution of the cells. If early shock were
to induce such local irregularities these could be significant and have later
repercussions only if in normal development the local ordering of the cells was
conserved. Proof that the local ordering of the cells is conserved through gastrulation, is perhaps the most remarkable, if least emphasized result to emerge from
vital dyeing experiments whose history stretches back over half a century (Vogt,
1929; Lovtrup, 1966,1975; Keller, 1975,1976). They show that dye marks made
prior to gastrulation become plastically deformed but not fragmented within
any one layer of the neurula. This indicates that topological integrity is maintained throughout gastrulation; the tissue redistribution of gastrulation occurs
without tearing such that cells originally close together within the same germ
layer of the blastula will be found (or their descendants will be found) still close
together in the neurula.
Current thinking on the performance of stress-free topological redistribution
during early development envisages the germ layers behaving as elasticoviscous
liquids in which the cells are individually deformable and free to slip on one
another (Phillips, Steinberg & Lipton, 1977; Phillips & Steinberg, 1978). If all
the cells slip a little in the course of plastic deformation of the tissue layers,
stresses are minimized and local neighbour relations minimally changed. Impairment of liquid behaviour would lead to the build up of stresses during
tissue redistribution released by tearing and the fault-like movement of one line
of cells on another parting old neighbours and making new neighbours of cells
previously at a distance.
It is necessary to consider therefore whether local alterations in neighbour
relations could be responsible for abnormalities in the segmental pattern.
Somitogenesis in amphibia. II
265
4. The kinematic nature of the waves associated with segmentation
A wave that depends upon the propagation of a signal will be stopped at a cut
across its path; a wave not so stopped that appears to jump across the cut is
termed a kinematic wave (Zeeman, 1975). A simple transection experiment
will therefore determine the nature of a wave of change crossing the embryo.
Embryos cut transversely in two after gastrulation and before any somites
have formed, and the two halves reared separately, segment the same number of
somites as intact embryos (Deuchar & Burgess, 1967). Pearson & Elsdale (1979)
have confirmed this result and determined that the somites are segmented in the
correct order, somite formation commencing in the posterior fragment immediately after it is completed in the anterior piece. Using temperature shocks they
also proved that in their experiments the prior wave had not reached the posterior halves at the time of their separation. A length of unsegmented paraxial
mesoderm cut out and replaced after rotation through 180° segments caudocranially Pearson (unpublished); the same result has been obtained in the chick
(Christ, Jacob & Jacob, 1974; Menkes & Sandor, 1977).
These results rule out a stimulus propagated from cell to cell along the axis in a
cranio-caudal direction responsible for the wave of rapid cell change, and imply
that the cells of the paraxial mesoderm behave autonomously according to
endogenous countdowns, the progress of the wave reflecting the layout of the
cells along the axis in the order in which they are pre-set to change. The transection experiments show that the countdowns are set not later than the end of
gastrulation. However, they are probably set much earlier for we know of no
evidence for regulation of the temporal course of somitogenesis after the blastula stage. Indeed Zeeman (1974) has suggested that the countdowns are set in
ordered sequence with the passage of a primary wave identified with mesodermal
induction in the blastula (Nieuwkoop, 1969). The maintenance of topological
integrity would be essential to ensure that the layout of the cells in the blastula
was correctly carried over into the neurula under these circumstances.
5. An hypothesis to account for abnormal somitogenesis following early shock
Each cell of the prospective paraxial mesoderm in the early gastrula is set to
undergo an abrupt change in behaviour at a later time. This change is one of the
components in the establishment of the segmental pre-pattern. The pre-somitic
cells in the early gastrula are arranged in a graded series with respect to the times
they are set to perform their changes. Under normal circumstances following
the redistribution of the tissue during gastrulation, starting in the late neurula
a sharp coherent frontier between the changed and unchanged cells sweeps
slowly, tailwards down the axis. This result depends on the maintenance of
topological integrity throughout gastrulation to ensure that the layout of presomitic cells along the axis in the neurula is a continuous transformation of the
layout of their precursors in the blastula. Temperature shock alters the prop-
266
T ELSDALE AND M. PEARSON
Table 5. Events leading up to somite formation and reaction to temperature shock
Stage
Blastula
Early gastrula
'Mid-gastrula'
Late gastrula onwards
Mid neurula
onwards
Development
Reaction to
temperature shock
Somite
specification
component
Affected
Induction of mesoderm.
Pregastrulation movements
Major pre-involutory
movement. Involution
Disturbed (but early part of
Wave
period not explored)
Movements disturbed. Local Wave
neighbour relations irreversibly disordered
Lesser and slower gastrulation Refractory. Local neighboui
movements
relations stable
Co-ordination of paraxial
Breakdown of co-ordination Co-ordifollowed by recovery
mesoderm cells
nation
None
Determined somites
(pre-pattern)
Formed somites
None
erties of the non-involuted prospective mesoderm in a way that makes gastrulation difficult, with the result that local breakdowns of topological integrity
occur with consequent re-ordering of local neighbour relations among the cells.
Across a region of tissue within which the order of the cells has been mixed up,
the wavefront of abrupt change fragments and a confused pre-pattern results.
In a concurrent study of the effects of early temperature shocks carried out on
Xenopus, Cooke (1978) proposed that early shocks might disturb development
of the molecular machinery of hypothetical cellular oscillators, the entrained
activities of which conferred co-ordination. This was an attempt to account for
the seemingly random distribution of abnormalities along the somite files following shocks at any time during the early sensitive period. Because Rana embryos
develop more slowly than Xenopus, and large numbers all at the same stage are
obtainable it has been possible to investigate the different stages of the early
sensitive period with a greater temporal resolution than was possible using
Xenopus. This has resulted in the recognition of the trends described in this
paper showing that the distribution of abnormalities is not in fact random but
correlated with the precise stage of gastrulation shocked. Similar trends have
since been confirmed in Xenopus (Cooke, personal communication). The data
presently available thus permit the more concrete hypothesis here presented.
Our view of the significant events leading up to somite formation and the
reactions to temperature shock are summarized in Table 5.
The authors wish to thank the Medical Research Council for occasional support to M. J.
Pearson during a prolonged period of unemployment, Jonathan Bard, Jonathan Cooke,
Duncan Davidson and Christopher Zeeman for their interest and helpful discussion, Allyson
Ross for technical assistance, Sandy Bruce for art work and photography and Elaine Smith
for manuscript preparation.
Somitenesis in amphibia. II
267
REFERENCES
CHRIST, B., JACOB, H. I., & JACOB, M., (1972). Experimentelle Untersuchungen zur Somiten-
entwicklung beim Huhnerembryo. Z. Anat. EntwGesch. 138, 82-97.
J. (1978). Somite abnormalities caused by short heat shocks to pre-neurula stages of
Xenopus laevis. / . Embryol. exp. Morph. 45, 283-294.
COOKE, J. & ZEEMAN, E. C. (1976). A clock and wavefront modelf for control of the
number of repeated structures during animal morphogenesis. /. theor. Biol 58, 455-476.
DEUCHAR, E. M. & BURGESS, A. M. C. (1967). Somite segmentation in amphibian embryos:
is there a transmitted control mechanism? / . Embryol. exp. Morph. 17, 349-358.
ELSDALE, T., PEARSON, M. & WHITEHEAD, M. (1976). Abnormalities in somite segmentation
following heat shock to Xenopus embryos. / . Embryol. exp. Morph. 35, 625-635.
KELLER, R. E. (1975). Vital dye mapping of the gastrula and neurula of Xenopus laevis. I.
Devi Biol. 42, 222-241.
KELLER, R. E. (1976). Vital dye mapping of the gastrula and neurula of Xenopus laevis. II
Devi Biol. 51, 118-137.
LOVTRUP, S. (1966). Morphogenesis in the amphibian embryo. Cell type distribution, germ
layers and fate maps. Acta zool. Stockh. 47, 209-276.
LOVTRUP, S. (1975). Fate maps and gastrulation in Amphibia-a. critique of current view.
Can. J. Zool. 53, 473-479.
MENKES, B. & SANDOR, S. (1977). Somitogenesis: regulation potencies, sequence determination, and primordial interactions. Vertebrate Limb and Somite Morphogenesis Ed. D. A.
Ede, T. R. Hinchliffe & M. Balls, pp. 405-420.
NIEUWKOOP, P. D. (1969). The formation of the mesoderm in Urodelan amphibians. 1.
Induction by the endoderm. Wilhelm Roux Arch. EntwMech. Org. 162, 341-373.
NIEUWKOOP, P. D. & FABER, J. (1956). Normal Table of Xenopus laevis (Daudin). Amsterdam:
North Holland.
PEARSON, M. J. & ELSDALE, T. (1979). Somitogenesis in Amphibia. I. Experimental evidence
for an interaction between two temporal factors in the specification of the osmite pattern.
/. Embryol. exp. Morph. 51, 27-50.
PHILLIPS, H. M., STEINBERG, M. S. & LIPTON, B. H. (1977). Embryonic tissues as elasticoviscous liquids. II. Direct evidence for cell slippage in centrifuged aggregates. Devi Biol. 59,
129-134.
PHILLIPS, H. M. & STEINBERG, M. S. (1978). Embryonic tissues as elasticoviscous liquid: I.
Rapid and slow shape changes in centrifuged cell aggregates. /. Cell Sci. 30, 1-20.
SHUMWAY, W. (1940). Stages in the normal development of Rana pipiens. I. External form.
Anat. Rec. 78, 139.
VOGT, W. (1929). Gestaltunganalyse am Amphibienkeim mit ortlicher Vitalfarbung. II. Teil.
Gastrulation und Mesodermbildung bei Urodelen und Anuren. Wilhelm Roux Arch.
EntwMech. org. 120, 384-406,
ZEEMAN, E. C. (1975). Primary and secondary waves in developmental biology. Lectures on
Maths in the life Sciences, 7. Amer. Math. Soc, Rhode Island, U.S.A. pp. 69-161.
COOKE,
{Received 9 February 1979, revised 19 August 1979).