Abortive Development in the Crested Newt Triturus Cristatus

Development 100, 65-72 (1987) 65 Printed in Great Britain © The Company of Biologists Limited 1987

Abortive development in the crested newt Triturus cristatus

H. WALLACE

Department of Genetics, University of Birmingham, PO Box 363, Birmingham BIS 2TT, UK

Summary

Crested newts exhibit a balanced lethal system which abortive embryo. causes development to abort in 50 % of tailbud em- Two possible modes of origin for the balanced bryos. The main features of arrested embryos are lethal system are discussed, either from ancestral described, shown to be constant in a range of culture autosomes or from sex chromosomes. The latter is conditions and found to be unchanged by parabiosis favoured because it involves fewer assumptions. The or extirpation of the central nervous system or heart. subsequent evolution of the system should inevitably Two categories of abortive embryos can be dis- lead to progressively earlier embryonic mortality by tinguished in some spawnings. A corresponding seg- an accumulation of multiple recessive lethal factors on regation of an easily identified marker chromosome each chromosome variant. confirms previous reports that chromosome 1 carries the recessive lethal factors. The marker chromosome Key words: developmental genetics, embryonic lethality, is tentatively associated with the blistered form of sex chromosomes, Triturus cristatus, newt.

Introduction and death in T.c. carnifex tailbud embryos, together with some experimental tests of whether or not a In one of the earliest descriptions of amphibian primary defect is localized in any particular tissue. development, Rusconi (1821) remarked that almost The data reported here were accumulated over the half the embryos of Triturus cristatus perished at the last three years, mainly while checking an anomalous tailbud stage. He concluded that the development of spawning that showed high embryonic survival. Most this species must be relatively sensitive to culture of the observations were obtained by breeding a conditions, as he encountered no such mortality second generation from this stock, which showed a among T. palmatus embryos. Spemann (1921), Gliick- regular 50% embryonic mortality and the corre- sohn (1932) and Stief (1940) also noticed a high sponding segregation of a distinctive marker chromo- mortality of T. cristatus embryos, without attempting some. to explain it. Macgregor & Homer (1980) redis- covered this embryonic mortality and established its genetic basis, as a consequence of finding that adult Materials and methods crested newts were invariably heterozygous for a heterochromatin pattern on the largest pair of All observations were made on T.c. carnifex, using six pairs purchased from Xenopus Ltd as a single importation and chromosomes. This chromosome 1 heteromorphism their descendents. Adult females were identified by the is apparently maintained by a balanced-lethal system pattern of ventral spots and isolated when laying. Paternity in natural populations of crested newts, including was uncertain for most spawnings, as a succession of males three subspecies of T. cristatus and the closely related were used to ensure successful courtship. Matings were T. marmoratus (Sims, Macgregor, Pellatt & Horner, induced by injecting chorionic gonadotrophin (100-200 i.u. 1984). Different culture conditions and temperature Pregnyl) in winter or occurred spontaneously in late spring. regimes have failed to alleviate the embryonic mor- Grass was provided for the spawning initially, so that eggs tality or to distinguish between the presumed two could be collected daily by cutting off folded grass blades. types of lethal homozygote (Wallace, 1984; Horner & This technique was later improved by substituting nylon Macgregor, 1985). ribbons for the grass, scraping the eggs off each ribbon daily and culturing batches of up to 20 eggs in a Petri dish. The The following account provides details of the eggs were then completely visible and less often damaged in developmental defects associated with retardation laying or collection. 66 H. Wallace

Before operations, the egg capsule was surface sterilized in 70 % ethanol for 15 s and removed with fine forceps in Steinberg solution (Rugh, 1962) containing penicillin and streptomycin. Extirpations and parabiosis were performed in the same medium on 2 % agar using tungsten needles (Popiela, 1976; Wallace, 1962). Embryos used for histology were also decapsulated, fixed in neutral formalin or Bouin, processed as 8/wn serial sections and stained either with haematoxylin and eosin or Feulgen and light green. Mitotic metaphases were accumulated by culturing decapsulated wounded embryos for up to 24 h in colchicine (0-5 ragml"1 Steinberg) in small plastic weighing dishes. The isolated heads of these embryos were then swollen in diluted medium for 15 min before fixation in methanol-acetic acid (3:1). Squashed preparations of the heads were subjected to a variety of C-banding techniques and stained with Giemsa (Gurr's R66). The technique described by Sims et al. (1984) provided the best results after brief fixation. Material stored in fixative at — 20 °C for 3-18 months only Fig. 1. Camera-lucida drawings of normal embryos responded to more severe alkaline extraction, 0-07 M- (A-D) aged 4, 5, 7 and 10 days at 20°C; Harrison stages barium hydroxide at 65°C for 30-60min. 24, 30, 35 and 38. Defective embryos (E-H) at the same ages show increasing retardation, dorsal blisters (E-G) and less regular pigmentation (G,H). Capsules shown in Results A and E are 5 mm long. Normal development Mated females typically lay 200-300 eggs during a onset of retardation even when comparing virtually period of 2-3 weeks. Even at the height of spawning, synchronous embryos. Several spawnings, however, they lay less than 40 eggs in one night and relatively have revealed an early manifestation of the defective few during the day. It is therefore more convenient to syndrome in head-process embryos (stage 24). Be- time development from the first cleavage, about 6h tween 10 % and 30 % of the embryos then carry after deposition at 20°C. Successive cleavages occur dorsal blisters over the brain and spinal cord at 90 min intervals. The ensuing morula (12-18 h) has (Fig. 1E-G). Almost all such embryos become ar- a corrugated animal hemisphere before the blastocoel rested, generally at an earlier tailbud stage than the enlarges dramatically at 22—24 h. Gastrulation occu- remaining defective embryos which lack blisters. A pies all the second day and neurulation all the third comparison of blistered specimens with contempor- day. The neurula first elongates by forming a head ary normal embryos suggests that at least some are process at 80-96 h (Harrison stages 21-26; Fig. 1A) already retarded at the head process stage. All before the tailbud becomes visible. Elongation of the defective embryos show subnormal tail growth, only trunk and tail then causes the embryo to become form the primordia of balancers and gills (Fig. 1H), arched within its capsule (stages 27-32; Fig. IB). and rarely show a heart beat, circulation or spon- Only half of these tailbud embryos show continued taneous movement. They are most easily dis- growth and an intensified pigmentation in two dorsal tinguished from younger normal embryos by their stripes, widely separated over the head (Fig. 1C,D). pigmentation, which covers the hindbrain instead of The heartbeat is usually evident by 6 days and blood separating into two dorsal stripes. Their development circulates through the growing gills by 7 days in these is virtually arrested after 7 days, but many survive for normal embryos which generally hatch at 9-12 days. most of the following week and become oedematous This description corresponds to the normal table before degenerating or succumbing to fungus. constructed by Horner & Macgregor (1985) using The main symptoms of retardation and tailbud stages derived from Xenopus and thus indirectly from arrest occur at all the temperatures tested, which Harrison stages. The latter were found suitable for approach the limits of embryonic survival. Uncleaved T. cristatus and extended to include larval life by eggs tolerate gradual warming to 25 °C, to cleave 5 h Gliicksohn (1932), and are reproduced in Rugh later and then at hourly intervals. The temperature (1962), so it seems preferable to use them here. can be safely increased to 28 °C only after gastrulation. In both cases, tailbud embryos become pig- Abnormal development mented by 4 days and half of them are clearly No abnormality is evident prior to the closure of the retarded at 5-6 days. Eggs chilled to 12°C cleave at neural tube (stage 20) and it is difficult to pinpoint the roughly 5h intervals, complete neurulation after 7 Balanced lethal factors in newts 67

Table 1. Summary of spawnings showing numbers of that the presence or absence of dorsal blisters dis- hatched larvae and arrested tailbud embryos after tinguishes the expected two types of lethal homozy- development at the indicated temperatures gote. Line Hatched Arrested

1 7 matings of purchased adults (12-25 °C) 352 333 (48 %) Experimental studies 2 1 mating as above: (a) at 20°C 50 52 (51 %) The possibility that much of the defective syndrome (b) at 25-28 °C 41 21 (34 %) 3 9 matings of survivors from line 2b: could result from failure to establish a blood circu- (a) at 20 °C 233 222 (49 %) lation (Wallace, 1984) has been tested in two ways. (b) at 24 °C 493 457 (48 %) Random pairs of embryos were joined in parabiosis at (c) at 28 °C 308 314 (50 %) stages 26-28. Only 21 twins out of 40 operations 4 8 matings of F] from purchased stock 317 293 (48 %) remained attached 2 days later, when defective em- (20 °C) 5 Samples from line 4 isolated as: bryos were already distinctly retarded. Six twins were (a) normal head process (hp) embryos 193 81 (21 %) both normal, ten were both defective and five con- (b) hp embryos with dorsal blisters 2 103 (27 %) sisted of a defective embryo attached to a normal one. Instead of achieving a circulation, the latter defective embryos accumulated particularly large days, and half the tailbud embryos are arrested by blood islands and two of them eventually showed a 14-16 days. Horner & Macgregor (1985) reported slow movement of erythrocytes in the posterior similar observations over the range 10-25°C. ventral vein. Although they survived longer than Development is not particularly sensitive to other arrested controls, even these two cases showed vir- culture conditions within this temperature range. The tually the same symptoms of tailbud arrest. Each had embryos apparently require only a minimal amount formed a tailfin and showed intensified pigmentation of water and thus access to oxygen. Some water is by the time their hosts were swimming actively and probably absorbed immediately after oviposition and ready to feed. more is needed at hatching, but normal development The second test involved extirpating the heart occurs in distilled water or the film of water held by a primordium from 24 unpigmented tailbud embryos folded grass blade, or even when entirely exposed to (stage 28). All the embryos survived this operation air in a humid chamber. Tailbud arrest occurs rou- and could be scored as 8 retarded and 16 normal tinely in about half the embryos exposed to these specimens 2 days later. The retarded ones became conditions. arrested and degenerated within 6 days. The normal ones developed surprisingly well in the absence of Table 1 summarizes virtually all the embryos a circulation, into rather stunted and oedematous reared, apart from samples removed for surgery or swimming larvae. They grew normal balancers, short fixation, by pooling data from different matings or gills and notch-stage arms, and formed typical pig- culture conditions that showed a similar mortality at ment stripes. Evidently, neither operation had a tailbud stages. Eggs from the original stock were marked influence on the genetic predisposition of cultured at several temperatures (lines 1 and 2) and these embryos. The only aspect of the defective only samples from one mating kept at high tempera- syndrome that could be a consequence of circulatory ture after gastrulation showed a significant departure 2 failure is the oedema that occurs in moribund ar- from 50% mortality (^ = 6-452; P<0-02). I could rested specimens. not repeat this mating, but reared 30 of the high An alternative prediction that much of the defec- survival Fx from line 2b to maturity and sibmated tive syndrome might be a secondary effect of earlier nine pairs. Their F2 progeny showed 50 % tailbud neural necrosis was tested by extirpating the brain mortality even at the highest tolerated temperature and trunk spinal cord from 26 head-process embryos (line 3). Another six Fj sibmatings and two involving at stages 21-25. The success of the operation was less-closely-related females (line 4) provided samples judged by the headless appearance and virtually from which embryos with dorsal blisters were isolated unpigmented trunk of all specimens. Retarded tail- at the head-process stage. The proportion of blistered bud specimens (12 cases) survived up to 10 days, but embryos varied greatly in the small samples exam- only one developed primordia of the gills and bal- ined, but the accumulated data shown in line 5 permit ancers. The remaining 14 cases were identified as a further inference. Dorsal blisters allow a fairly normal by the growth of the unoperated tail with its accurate prediction of developmental arrest and typical fin and pigmentation. These embryos all identify about half of the embryos that will eventually developed a rapid circulation, long gills and bal- abort. These two classes of defective embryos each ancers, and the best cases had arms with three digits comprise about 25 % of the total, so it is possible after 10 days. Since the operation neither affected 68 H. Wallace tailbud arrest nor created a phenocopy of the defec- arrest occur before overt cellular differentiation of tive syndrome in genetically viable embryos, develop- most tissues. Some arrested embryos possess an optic mental arrest must occur independently of the neural cup containing a small lens vesicle without fibres, necrosis detailed in the following section. others possess an optic lobe without any lens. The notochord is vacuolated, however, contractile muscle differentiates in the somites, and both melanophores Histology and erythrocytes havebecome pigmented. Serial sections of abnormal and control embryos at Little further differentiation occurred when defec- the stages shown in Fig. 1 revealed that the most tive embryos were maintained beyond their normal striking defect concerns the central nervous system lifespan in parabiosis. Disorganized regions of white and head mesenchyme. Massive cell death with matter formed in the brain, suggesting a chaotic pycnotic nuclear fragments and cellular debris occurs growth of nerve fibres. No jaw or gill cartilage could in these tissues of head-process and early tailbud be identified except as extensions of the host visceral embryos identified by the presence of dorsal blisters, skeleton. Nuclei of most tissues in both normal and and also in some embryos that appeared normal by defective specimens contained one to four nucleoli. external criteria. The blisters are packed with cellular debris, probably derived from the adjacent neural Chromosomes tube and crest. Parts of the hindbrain and anterior Retarded tailbud embryos have a greatly reduced spinal cord have usually collapsed, leaving patches of incidence of cell division. Even after colchicine treat- debris in the remaining cavities. Later tailbud em- ment, preparations contained only 0-20 brain meta- bryos invariably show collapsed regions of the central phases and excessive numbers of pycnotic nuclear nervous system and persistent cell debris, although fragments, whereas preparations from normal tailbud pycnotic nuclei are now less common (Fig. 2). Oc- embryos usually provided more than 100 metaphases. casional pycnotic nuclei are also present in the heart, Colchicine treatment of head-process stages retarded pronephros, notochord and somites, and probably in development enough to impede the detection of all endodermal derivatives, but cell death does not defective embryos and produced some pycnotic nu- affect the integrity of these organs. Retardation and clei in all preparations, which could still be scored

Fig. 2. (A) Section through the cranial flexure, eyecups and lens of a normal tailbud embryo at stage 35. (B) Oblique section in the same region of a contemporary defective embryo, showing one optic lobe without a lens (left) and collapsed hind brain (upper centre). A mass of necrotic tissue lies above the first gill pouch on the right. Scale bar, 0-5 mm. Balanced lethal factors in newts 69 unambiguously as having abundant or scarce meta- them could be scored with confidence, but they all phases and assigned to viable or defective embryos on carried the marker chromosome. These observations that basis. lead to the tentative conclusion that the marker Chromosome 1 is easily recognized after C-banding chromosome is responsible for precocious retardation by the heterochromatic segment which occupies most and dorsal blisters in Ad/Ad and Ad/Ap defective of its long arm, sometimes interrupted by a pale gap embryos. and typically leaving a pale tip. The basis of the heteromorphism is a series of intensely dark bands within the heterochromatic segment. This banding Discussion pattern changes with chromosome contraction and is obscured by overstaining, as well as exhibiting natural The observations reported here fully support the variation between individuals. Chromosome 1A has genetic nature of the embryonic mortality, partly by been recognized as having a generally darker het- excluding a wide range of environmental influences erochromatic segment or one more heavily decorated and partly by showing it corresponds to a chromo- with intense bands than that of IB. The idealized some segregation analogous to that originally de- diagrams provided by Macgregor & Horner (1980) scribed by Macgregor & Horner (1980). The develop- and Sims et al. (1984) differ considerably from each mental retardation is associated with cell death in other and from Fig. 3A in this respect, leaving some several tissues, most strikingly in neural derivatives doubt over the objectivity and reliability of the where cell division and yolk utilization suggest meta- distinction. In contrast to the pale-tipped long arms bolic activity is highest. Simple experiments indicate typical of chromosome 1 (hereafter Ap and B), that the retardation and arrest are not caused by a however, there is a variant with an almost entirely single local defect in either the nervous or the circu- dark tip (Ad) which is much more easily identified latory system. Consequently, the defect is best (Fig. 3; cf. Macgregor, Horner & Sims, 1983). characterized as an autonomous cell lethal in Table 2 shows how this marker chromosome segre- Hadorn's (1961) terminology. That implies the failure gated in samples of embryos diagnosed as viable or of some general metabolic pathway which becomes defective by the criteria stated above. The observed critical in any cell once the relevant maternal reserves classes are entirely consistent with there being com- become exhausted. We can probably exclude the two plete linkage between the lethal system and the most obvious examples of reserves stored in the marker on chromosome 1, as shown especially by the oocyte, ribosomal RNA and histones. The clustered single class of viable embryos in line 1 of Table 2. arrangement of these highly repeated genes is hardly Line 2 is more complicated because obviously hetero- compatible with involvement in a balanced lethal morphic pairs of chromosome 1 occur in approxi- system, as argued later, and the major sites of the mately half the viable and defective embryos. The clusters are located on other chromosomes (Morgan, expected Mendelian ratios are certainly disturbed in Macgregor & Colman, 1980; Gall, Stephenson, Erba, Table 2 by the absence of scorable metaphases from Diaz & Barsacchi-Pilone, 1981). The presence of up some defective embryos. Furthermore the inferred to four nucleoli in cells of defective embryos is a fair genotypes are partly based on diagnosis of the whole indication of ribosome biosynthesis, but not that the embryos and the abundance of metaphases, as Ap ribosomes must be functional. Failure to produce one and B could not often be distinguished with any or other of the ribosomal proteins remains a possi- confidence. bility, for instance, which would account for the Those defective embryos that carried the marker considerable resemblance between the defective syn- chromosome Ad had appreciably fewer metaphases drome and that of the anucleolate mutant of Xenopus than those which did not (BB). Each of the samples (Wallace, 1960, 1962). used in Table 2 included some head-process or tail- One feature to attract previous comment is that the bud embryos with dorsal blisters. Only one sample two lethal homozygotes are virtually indistinguish- was isolated on this basis and only five (out of ten) of able. Two independent mutations could produce Table 2. Numbers of embryos scored according dark (D) or pale (P) tips of chromosome 1

Viable embryos Defective embryos

Line Inferred parental genotype D/P P/P D/D D/P P/P

1 Ad/B x Ad/B 32 0 13 0 11 2 Ad/B x Ap/B 9 12 0 5 4

Inferred genotype Ad/B Ap/B Ad/Ad Ad/Ap B/B 70 H. Wallace identical effects by coincidence, of course, especially gene complex so as to produce complementary if a variety of maternal reserves are exhausted by the lesions in a single metabolic pathway (Macgregor & head-process stage. Alternatively, a translocation Horner, 1980; Sims et al. 1984). Such complemen- between chromosomes 1A and IB might have split a tation in the heterozygote implies the mutations

Ad Ap

3A B .w^ r.

Fig. 3. (A) Banding pattern of chromosome 1 variants. (B) Metaphase from a neurula showing chromosome 1 heteromorphy: Ad (arrow) and either Ap or B (open arrow). Scale bar, 10\ixa. (C-E) Chromosome 1 pairs of tailbud embryos from a single spawning, identified as explained in the text. (C) Ad/B from a viable embryo; (D) Ad/Ad and (E) B/B both from defective embryos. Balanced lethal factors in newts 71 concern functionally related genes rather than ident- on its bearer and hence tend to spread, it seems that ical copies of a repeated sequence. The evidence the long term stability of the system actually requires presented here that it is possible to distinguish the the presence of multiple recessive lethals on both A two types of lethal homozygote, at least in some andB. matings, suggests that at least two independent lethal The many visible variants of chromosome 1A, first mutations have occurred and probably more than that detected by Callan & Lloyd (1960), surely reflect in some chromosome variants. underlying genetic differences and, according to this Sims et al. (1984) pointed out that this balanced analysis, probably include a diversity of recessive lethal system is one of considerable antiquity and is lethal factors. The immediate cause of death in the associated with several peculiar features of chromo- homozygotes is merely the mutation that has the some 1. The heterochromatic segments contain un- earliest effect, so that random mutation results in an usual amounts of highly repeated DNA sequences inevitable evolution to embryonic mortality in the and are achiasmate regions where no recombination period when maternal reserves become exhausted can occur in either sex. They envisaged an evolution and need to be replaced by zygotic gene activity. beginning with a chromosomal rearrangement which Other cases of natural balanced lethal systems, prevented recombination in this segment, followed Oenothera (Cleland, 1972) and Ellobius (Lyapunova, by a recessive lethal mutation on both 1A and IB and Vorontsov & Zakarjan, 1975), also show 50 % embry- finally the accumulation of repeated sequences in onic mortality. The former has complex chromosome both the genetically isolated segments. Each of these rearrangements and the latter involves sex chromo- steps requires that a novel mutation should spread to somes. I suspect both may have played a part in the fixation in the ancestral population, either for no abortive development of crested newts. apparent reason or despite a strong selective disad- vantage. I am grateful to Professor J. C. Beetschen for referring All the peculiarities of chromosome 1 mentioned me to Rusconi's pioneering study, to several colleagues for above are common characteristics of sex chromo- providing their recipes for C-banding, and to Sir Kenneth somes, which alerted me to the following alternative Mather for genetic counselling. scheme for the origin of the balanced lethal system. If 1A and IB were the sex chromosomes of the ancestral References species, they would be expected to already contain a differential segment where no recombination oc- CALLAN, H. G. & LLOYD, L. (1960). Lampbrush curred in the heterogametic sex AB and consequently chromosomes of crested newts Triturus cristatus where recessive lethal mutations had already accumu- (Laurenti). Phil. Trans. Roy. Soc. Lond. 243B, lated in the chromosome (say B) restricted to that 135-219. sex. The single critical event is that this system of sex CLELAND, R. E. (1972). Oenothera: Cytogenetics and determination should be supplanted by another - Evolution. New York: Academic Press. presumably the present XY system on chromosome 4 GALL, J. G., STEPHENSON, E. C, ERBA, H. F., DIAZ, M. (Sims et al. 1984) - which only operated effectively in O. & BARSACCHI-PILONE, G. (1981). Histone genes are the former heterogametic sex AB. Then the superior- located at the sphere loci of newt lampbrush ity of the XY sex determination should drive AB to chromosomes. Chromosoma 84, 159-171. , fixation, owing to the relative infertility of A A and GLUCKSOHN, S. (1932). Aussere Entwicklung der the pre-existing lethality of BB. It is certain that such Extremitaten und Stadien einteilung der Larvenperiode von Triton taeniatus Leyd. und Triton cristatus Laur. an event has occurred at least once in the ancestry of Wilhelm Roux Arch. EntwMech. Org. US, 341-405. european newts. Past glaciations could have provided HADORN, E. (1961). Developmental Genetics and Lethal suitable evolutionary pressure for an improved means Factors. London: Methuen. of sex determination, as the distorted sex ratios HORNER, H. A. & MACGREGOR, H. C. (1985). Normal recorded previously (Wallace, 1984) were almost development in newts (Triturus) and its arrest as a certainly caused by low temperature. consequence of an unusual chromosome situation. J. Muller's (1918) analysis of balanced lethal factors Herpetol. 19, 261-270. explains the subsequent evolution of this system, LYAPUNOVA, E. A., VORONTSOV, N. N. & ZAKARJAN, G. however it originated. Recessive lethal mutations G. (1975). Zygotic mortality in Ellobius lutescens (Rodentia: Microtinae). Experientia 31, 417-418. should accumulate on both A and B, as the only MACGREGOR, H. C. & HORNER, H. (1980). restriction to that process is the presence of a func- Heteromorphism for chromosome 1, a requirement for tional allele on one chromosome to complement each normal development in crested newts. Chromosoma 76, recessive lethal on the other. Since back-mutation of 111-122. the single recessive lethal on either chromosome MACGREGOR, H. C, HORNER, H. A. & SIMS, S. H. (1983). would immediately confer a 50 % selective advantage Newt chromosomes and some problems in evolutionary 72 H. Wallace

cytogenetics. In Kew Chromosome Conference II (ed. P. heteromorphism and independent chromosome E. Brandham & M. D. Bennett), pp. 283-294. evolution. Chromosoma 89, 169-185. London: Allen & Unwin. SPEMANN, H. (1921). Die Erzeugung tierischer Chimaren MORGAN, G. T., MACGREGOR, H. C. & COLMAN, A. durch heteroplastiche embryonale transplantation (1980). Multiple ribosomal gene sites revealed by in zwischen Triton cristatus und taeniatus. Wilhelm Roux situ hybridisation of Xenopus rDNA to Triturus Arch. EntwMech. Org. 48, 533-570. lampbrush chromosomes. Chromosoma 80, 309-330. STIEF, C. (1940). Untersuchungen uber die Formbildung MULLER, H. J. (1918). Genetic variability, twin hybrids der Hautdrusen bei Amphibien mit Hilfe and constant hybrids, in a case of balanced lethal heteroplastischer transplantation zwischen Triton factors. Genetics 3, 422-499. taeniatus und cristatus. Wilhelm Roux Arch. EntwMech. POPIELA, H. (1976). In vivo limb tissue development in Org. 140, 495-524. the absence of nerves: a quantitative study. Expl WALLACE, H. (1960). The development of anucleolate Neurol. 53, 214-226. embryos of Xenopus laevis. J. Embryol. exp. Morph. 8, RUGH, R. (1962). Experimental Embryology. Minneapolis: 405-413. Burgess. WALLACE, H. (1962). Prolonged life of anucleolate RUSCONI, M. (1821). Amours des salamandres aquatiques et Xenopus tadpoles in parabiosis. J. Embryol. exp. developpement du tetard de ces salamandres depuis I'oeuf Morph. 10, 212-223. jusqu'a Vanimal parfait. Milan: Paolo Emilio Giusti. WALLACE, H. (1984). Chromosome 1 and sex SIMS, S. H., MACGREGOR, H. C, PELLATT, P. S. & determination of crested newts. Caryologia 37, 3-8. HORNER, H. A. (1984). Chromosome 1 in crested and marbled newts (Triturus): an extraordinary case of {Accepted 16 January 1987)