Development of the tectum in Gymnophiones, with …ib.berkeley.edu/labs/mwake/papers/144.pdfDevelopment of the Tectum in Gymnophiones, With Comparison to OtherAmphibians ANDREA SCHMIDT1

Development of the Tectum in Gymnophiones, With Comparisonto Other Amphibians

ANDREA SCHMIDT1 AND MARVALEE H. WAKE2*1University of Bremen, Brain Research Institute, 28334 Bremen, Germany2Department of Integrative Biology and Museum of Vertebrate Zoology,University of California, Berkeley, California 94720-3140

ABSTRACT Tectal development in a number of caecilian (Gymnophiona:Amphibia) species was examined and compared with that in frogs andsalamanders. The caecilian optic tectum develops along the same rostrocau-dal and lateromedial gradients as those of frogs and salamanders. However,differences exist in the time course of development. Our data suggest that, asin salamanders, simplification of morphological complexity in caecilians isdue to a retardation or loss of late developmental stages. Differences in thetime course of development (heterochrony) among different caecilian speciesare correlated with phylogenetic history as well as with variation in lifehistories. The most pronounced differences in development occur between thedirectly developing Hypogeophis rostratus and all other species examined. Inthis species, the increase in the degree of morphological complexity is greatlyaccelerated. J. Morphol. 236:233–246, 1998. r 1998 Wiley-Liss, Inc.

KEY WORDS: caecilian species; late developmental stages; morphological complexity

The visual system of amphibians isstrongly correlated with phylogenetic his-tory (Roth et al., ’83, ’90; Wake, ’85; Schmidtand Wake, ’91). Differences in the visualsystem are apparent among the three ordersof amphibians (frogs [Anura], salamanders[Urodela], and caecilians [Gymnophiona]),as well as among species within each order.These differences concern the absolute andrelative sizes of eyes (Wake, ’85; Roth et al.,’90), eye morphology (Wake, ’85; Linke et al.,’86), degree of frontal direction of the eyes(Roth, ’87), patterns of visual projections(Fritzsch, ’80; Fritzsch et al., ’85; Himstedtand Manteuffel, ’85; Rettig and Roth, ’86),and tectal morphology (Roth et al., ’90, ’93).The morphologically most complex optic tec-tum among amphibians is found in frogs.They possess a multiply laminated tectum(Potter, ’69) in which many cells can be foundin the superficial neuropil. Such cells havemigrated from their periventricular originduring the course of development, and areknown as ‘‘migrated’’ cells. This condition isin contrast to salamanders and caecilians,which have a secondarily simplified tectalmorphology (Schmidt and Wake, ’91; Roth etal., ’94) with a more or less homogeneousperiventricular gray and a superficial neuro-pil that exhibits very few migrated cells.

Despite variable reduction of the visual sys-tem among caecilians, the degree of morpho-logical complexity (lamination and numberof migrated cells) in the most developed cae-cilian tectum is greater than in that in mostsalamanders (Schmidt and Wake, ’91, ’97),suggesting that internal dynamics might bemore important for tectal differentiationthan external influences. Compared withsalamanders, caecilians vary especially inthe localization of migrated cells. Althoughsalamanders possess very few migrated cellsthat are present equally in both the medialand lateral parts of the tectum, the majorityof migrated cells in caecilians occurs in thelateral part of the tectum (Schmidt andWake, ’91, ’97).

Many studies (e.g., Roth et al., ’93; Schmidtand Roth, ’93) have demonstrated that differ-ences in brain morphology are related toheterochrony (i.e., dissociations or relativechanges in patterns of development). For

Contract grant sponsor: Deutsche Forschungsgemeinschaft;Contract grant number: Schm 833/3-2; Contract grant sponsor:National Science Foundation.

*Correspondence to: Marvalee H. Wake, Department of Integra-tive Biology, 3060 VLSB, University of California, Berkeley, CA94720-3140; E-mail: [email protected]

JOURNAL OF MORPHOLOGY 236:233–246 (1998)

r 1998 WILEY-LISS, INC.

example, in salamanders, the simplificationof the tectum mesencephali is thought to bedue to paedomorphosis characterized by aloss or retardation of late ontogenetic pro-cesses, particularly the migration of cellsinto the superficial neuropil (Schmidt andRoth, ’93). It is not known whether the lowdegree of morphological differentiation inthe tectum of caecilians also is due to paedo-morphosis, or if other factors, such as thereduction of the peripheral visual system,might contribute to the simplification.

Two major factors have been shown to berelated to interspecific variation during braindevelopment: (1) differences in genome size(Roth et al., ’93), and (2) differences in lifehistories (Rettig and Roth, ’86; Schmidt etal., ’88; Wake and Roth, ’89). Both wereshown to play an important role during braindevelopment in salamanders (Schmidt andBoger, ’93; Roth et al., ’93).

In salamanders, an increase in genomesize is thought to be correlated with theoccurrence of paedomorphosis, in that itleads to a general slowing down of develop-ment (Wake, ’66; Cavalier-Smith, ’78; Hor-ner and MacGregor, ’83; Sessions and Lar-son, ’87; Roth et al., ’93, ’94). Frogs possesssmaller genomes than salamanders (Olmo,’83). The genome sizes of caecilians lie be-tween those of frogs and salamanders (M.H.Wake and S.K. Sessions, unpublished obser-vations). Presuming that there is a relation-ship between an increase in genome size andthe degree of paedomorphosis, one wouldexpect that paedomorphosis does not affectbrain development as strongly in caeciliansas it does in salamanders.

Roth et al. (’93) presented a phylogeneticanalysis of the evolution of the brain invertebrates, using 23 neuroanatomical char-acters, including those of the tectum. Theirmost parsimonious conclusion is that a com-plex brain is the ancestral condition, andsecondary simplification occurred indepen-dently in hagfishes, lepidosirenid lungfishes,and salamanders and caecilians. Therefore,assuming that frogs possess the ancestralcondition of adult tectal morphology and oftectal development, and given the informa-tion about the ontogenetic sequence of tectaldevelopment in frogs (Schmidt and Roth,’93), we have a means of evaluating whetherdifferences in the degree of morphologicalcomplexity in caecilians are also due tochanges in the time course of development,as has been shown in salamanders (Schmidt

and Roth, ’93) by comparing the tectal devel-opment in frogs with that of caecilians. Het-erochronies associated with differences inlife histories may be due to changes in theaction of hormones (Gilbert, ’91). For in-stance, experimental studies on skull devel-opment (Hanken and Summers, ’88a,b; Han-ken and Hall, ’88) and on the development ofthe nervous system (Schmidt and Boger, ’93)provide evidence that direct development isdue to an increase in thyroxine levels. Theconsequence is an acceleration of develop-mental processes as well as a decoupling ofontogenetic processes leading to the develop-ment of a mosaic of characters similar tothose found in many direct-developing sala-manders.

In order to investigate the relationshipamong different life history strategies andchanges in the time course of development,we compared the ontogenetic sequence oftectal development in a diversity of caecilianspecies that exhibit various life historymodes.

MATERIALS AND METHODSSpecies and stages examined

Tectal development was studied in fivespecies of three of the six families of theOrder Gymnophiona (Duellman and Trueb,’86; Nussbaum and Wilkinson, ’89; Hedgeset al., ’93); we currently accept Typhlonecti-dae as a distinct family, as do most research-ers on caecilian biology) that have differ-ent life history strategies. Families andspecies include Ichthyophidae: Ichthyophiskohtaoensis (Southeast Asia [Thailand];oviparous); Caeciliidae: Dermophis mexica-nus and Gymnopis multiplicata (both Cen-tral America [Mexico to Panama and CostaRica, respectively]; both viviparous); Hypo-geophis rostratus (Seychelle Islands; directdevelopment); and Typhlonectes compressi-caudus (South America [north-central]; vi-viparous). Most animals were collected, dis-patched, and preserved immediately in thefield.

Tectal development was studied in theheads of animals at different ontogeneticstages. Parameters for choosing stages were(1) presumed time past fertilization, and(2) size (total length [TL]): Ichthyophiskohtaoensis (newly hatched larva/66 mm);Dermophis mexicanus (1.5 months/20 mm, 2months/37 mm, 2.5 months/40 mm, 3months/49 mm, 5 months/79 mm); Gymno-pis multiplicata (2 months/30 mm, 3months/54 mm, 5 months/84 mm); Typhlo-

234 A. SCHMIDT AND M.H. WAKE

nectes compressicaudus (approximately 3months/46 mm); Hypogeophis rostratus (ageundeterminable/38 mm).

MethodsSpecimens were fixed in 10% neutral buff-

ered formalin for a minimum of 72 hr andpreserved in 70% ethanol. Heads were re-moved, decalcified in 3% formic acid, embed-ded in paraffin, and transversely sectionedat 10 µm. Every third slide was stained withpicroponceau (for connective tissue; nuclei ofbrain cells were particularly well stained)hematoxylin and eosin (H&E) (for contrast),or Mallory’s azan (for tissue distinctivity)according to standard procedures (Huma-son, ’79). Other series were stained withPalmgren’s silver specifically to display neu-ron morphology. In sections stained usingthese methods, regions of notable cell prolif-eration can be distinguished from those withno or low proliferation by dark staining ofnuclei and the appearance of elongated neu-rons in addition to compact cells.

RESULTSTectal development in frogs

and salamandersThe anuran tectum is a conspicuously mul-

tilayered structure that is divided into ninealternating cellular and fiber layers (Potter,’69). Many migrated cells (#30%) are foundin the superficial fiber layer. By contrast, thesalamander tectum is characterized by athick periventricular cellular layer and anouter fiber layer that contains only few mi-grated cells (Roth et al., ’93).

Rana temporariaThe optic tectum of the frog Rana tempo-

raria develops along rostrocaudal and latero-medial gradients. The lateral part of therostral tectum differentiates first and themedial part of the caudal tectum differenti-ates last (Schmidt and Roth, ’93). At earlystage 24 (Gosner, ’60), cell proliferation takesplace over the entire width of the ependymallayer, more strongly laterally than medially(Fig. 1A). A superficial neuropil is absent.This mode of cellular proliferation is foundover all the tectum (rostral, central and me-dial). From stage 25 (in the rostral tectum)and stage 31 (in the caudal tectum), cellproliferation occurs in both a medial and in alateral zone, with an intermediate zone oflow cell proliferation lying between thesezones. At this time, cellular proliferation is

somewhat more intense in the medial thanin the lateral zone. The superficial neuropilbegins to develop, and the first migratedcells appear. The lateral part of the tectum isalways the first part in which a superficialneuropil and migrated cells are found, andthe region where, compared with the medialzone, cell proliferation increases initially andthen also decreases first. From stage 30 (inthe rostral tectum) and stage 37 (in thecaudal tectum), the zone of medial cell prolif-eration decreases. Lamination starts at stage28 in the rostral tectum and at stage 37 inthe caudal tectum.

Pleurodeles waltlAs in Rana temporaria, the tectum of the

salamander Pleurodeles waltl develops alongrostrocaudal and lateromedial gradients. Themode of cellular proliferation is similar tothat of Rana temporaria. However, in gen-eral, cell proliferation in Pleurodeles is lessextensive than in Rana temporaria. Thisbecomes most obvious at midlarval and lateontogenetic stages. At early stages (33–37[staging according to Gallien and Durocher,’57], in the rostral tectum), cell proliferationoccurs all over the entire width of the epen-dymal layer, with a slight concentration inthe lateral part of the tectum (Fig. 1B). Thesuperficial neuropil begins to develop at stage34. The small numbers of migrated cells donot appear before stage 36. The tectum ofPleurodeles exhibits no distinct multiplelamination. Starting at stage 38, cell prolif-eration decreases, beginning in the rostraltectum. At this time there is a stronger cellproliferation in the medial than in the lat-eral zone of the tectum. Cell proliferation inthe lateral zone is minimal and also de-creases in the medial zone, starting in therostral tectum at stage 42. At this stage,Pleurodeles is only 20% throughout its lar-val period.

Tectal development in caeciliansThe caecilian tectum, like that of sala-

manders, appears rather simple and con-sists primarily of a periventricular cellularlayer and a superficial fiber layer, that, how-ever, contains relatively more migrated cellsthan salamanders. In caecilians, migratedcells occur primarily in the lateral part ofthe tectum rather than in the medial part.Among caecilians, there are differences inthe complexity of the cellular layer, as wellas in the number of migrated cells (Schmidt

CAECILIAN TECTAL DEVELOPMENT 235

and Wake, ’97). In some species (e.g., Gymno-pis multiplicata), the periventricular cellu-lar layer is homogeneous, whereas in otherspecies (e.g., Dermophis mexicanus, Ichthyo-phis kohtaoensis, Hypogeophis rostratus, andTyphlonectes compressicandus), it is tra-versed by fibers that in most cases do notconstitute continuous layers. Among the cae-cilian taxa examined, Hypogeophis, Typhlo-

nectes, and Ichthyophis possess a moderatedegree of morphological complexity and amoderate number of migrated cells withinthe superficial neuropil. Gymnopis and Der-mophis both possess many migrated cells.However, in contrast to Gymnopis, whichhas a homogeneous cellular layer, Dermo-phis has the most complex tectum amongthe genera examined.

Fig. 1. Sequence of tectal development in the frogRana temporaria (A,C,E) and the salamander Pleurode-les waltl (B,D,F). Transverse sections of the centraltectum. A,B: Early stages, shortly after hatching. C,D:Midlarval stages. E,F: Shortly before metamorphosis.Bar 5 100 µm. In both species, cell proliferation (ar-

rows) occurs all over the tectum during early stages(A,B) and is restricted to a lateral and medial zone atlater stages (C,D). Cell proliferation in Pleurodeles isless than in Rana and stops earlier in relative to thetime of metamorphosis.


In all the caecilian species examined, theontogenetic sequence of cellular prolifera-tion and cellular migration is basically thesame as that described for frogs. However,there are changes in the time course of devel-opment and in the intensity of cell prolifera-tion. As in Rana, cell proliferation takesplace over all the ependymal layer early indevelopment, with a concentration in thelateral part of the tectum (Figs. 2, 3, 6). Atlater developmental stages the rate of lat-eral cell proliferation decreases. In contrastto Rana, there is no increase in cell prolifera-tion in the medial zone. Tectal developmentin caecilians also proceeds along rostrocau-dal and lateromedial gradients. The lateralpart of the rostral tectum is the first part inwhich a superficial neuropil and migratedcells are found; the medial part of the caudaltectum is the region in which these pro-cesses occur last.

Dermophis mexicanus20-mm TL. Cell proliferation occurs over

the entire width of the ependymal layer inthe rostral, central, and caudal tectum (Fig.2A,C,E).Asmall amount of superficial neuro-pil that diminishes rostrocaudally is foundin the lateral tectum. There are no migratedcells.

37-mm TL. Cell proliferation is concen-trated in the lateral tectum. It is least in therostral tectum and most extensive in thecaudal tectum. Compared with the 20-mmindividual, cell proliferation in the 37-mmTL animal is decreased medially in all partsof the tectum. In the rostral and in thecentral tectum, a few migrated cells occur inthe ventralmost part of the lateral tectum.Substantial numbers of migrated cells arefound in the adjacent thalamic and tegmen-tal areas. No migrated cells are found in thecaudalmost part of the tectum. The superfi-cial neuropil is present in the lateral tectum,but there is very little neuropil in the medialtectum. The amount of superficial neuropildiminishes rostrocaudally.

49-mm TL. The rostral tectum lacks dis-tinctive zones of cellular proliferation (Fig.2B). In the central tectum, cell proliferationis restricted to a lateral zone (Fig. 2D). Thereis no cell proliferation in the medial zone.The caudal tectum (Fig. 2F) exhibits an ex-tensive cell proliferation that is concen-trated in the lateral zone. Less cell prolifera-tion occurs in the medial tectum. In therostral and central tectum, few migrated

cells are found in the lateral part. The cau-dal tectum does not have any migrated cells.The superficial neuropil is best developed inthe rostral tectum and diminishes caudad.The caudalmost tectum has only a smallamount of neuropil in both the lateral andthe medial areas (Fig. 2F).

79-mm TL. The rostral and the centraltectum lack zones of cellular proliferation.An extensive cell proliferation is found onlyin the caudal tectum where it is concen-trated laterally. Migrated cells are found inthe rostral as well as in the central and thecaudal tectum, but they are mainly re-stricted to the lateral tectum. Very few mi-grated cells occur in the medial tectum. Thesuperficial neuropil is most developed in therostral and central portions of the tectum,but a lesser amount of neuropil is present inthe caudal tectum, where it is better devel-oped in the lateral than in the medial part.Gymnopis multiplicata

30-mm TL. At this stage, tectal morphol-ogy is similar to that of the 37-mm specimenof Dermophis. There is little cell prolifera-tion in the rostral tectum (Fig. 3A), but itoccurs equally in the medial and the lateralzones. Extensive cell proliferation occurs lat-erally and to a lesser degree medially withinthe central tectum (Fig. 3C). Cell prolifera-tion increases caudad, and is greater in thelateral than in the medial zone (Fig. 3E).Very few migrated cells are apparent in therostral and central tectum, and none is foundin the caudalmost tectum. The superficialneuropil is best developed in the lateral partof the rostral tectum and diminishes caudo-medially.

54-mm TL. At this stage, there is little tono cellular proliferation in the rostral andcentral tectum (Fig. 3B,D). Extensive prolif-eration occurs in the lateral zone of thecaudal tectum (Fig. 3F), and less prolifera-tion occurs in the medial zone. Few migratedcells are found in the rostral tectum. Moremigrated cells are found in the central tec-tum and especially in the adjacent tegmen-tum. There are still no migrated cells in thecaudalmost tectum. The superficial neuropilis best developed in the lateral part of therostral tectum and diminishes caudomedi-ally.

84-mm TL. Cellular proliferation is com-pleted in the rostral, central, and caudaltectum. In all parts of the tectum (rostral,central and caudal), the superficial neuropil


is well developed, and a substantial numberof migrated cells appear in the lateral part ofthe tectum. Very few migrated cells are ap-parent in the medial part of the tectum.

Typhlonectes compressicaudus46-mm TL. Compared with other taxa,

tectal cell proliferation in Typhlonectes ismuch reduced. Hardly any proliferativezones can be distinguished except in thecaudal tectum (Fig. 4E), where cell prolifera-

tion occurs over the entire width of the epen-dymal layer (Fig. 4C). No difference in prolif-erative activity is found between the lateraland medial tectum as described in Gymno-pis and Dermophis. There are very few mi-grated cells in the lateral part of the rostraltectum (Fig. 4A). Compared with Dermophisand Gymnopis at similar sizes, the relativesize of the neuropil is very small. Also inTyphlonectes, the neuropil is more developedin the rostral than in the caudal tectum.

Fig. 2. Sequence of tectal development in Dermophismexicanus (viviparous). A,C,E: 20-mm animal. B,D,F:49-mm animal. Transverse sections of the rostral (A,B),central (C,D), and caudal (E,F) tectum. In the 20-mmanimal, cell proliferation (arrows) occurs all over theependymal layer in the rostral (A), central (B), and

caudal (C) tectum. In the 49-mm animal, this earlypattern of cell proliferation only occurs in the caudaltectum. In the central tectum, cell proliferation is re-stricted to a lateral zone. Proliferation already stoppedin the rostral tectum. Bar 5 100 µm.


Hypogeophis rostratus38-mm TL. At this stage, tectal develop-

ment is nearly complete. Neither rostral (Fig.4B), central (Fig. 4D), nor caudal tectum(Fig. 4F) possesses distinctive zones of cellu-lar proliferation. The number of migratedcells is very low within the rostral tectumbut increases caudad. Most migrated cellsare concentrated in the lateral tectum andadjacent tegmentum. Very few occur in the

medial part. The superficial neuropil is welldeveloped in the rostral, central, and caudaltectum.

Ichthyophis kohtaoensis66-mm TL. At this stage, tectal develop-

ment is well advanced. No distinctive zonesare found in the rostral, central, and caudaltectum (Fig. 5). A few migrated cells occur inthe rostral, central and caudal tectum. These

Fig. 3. Sequence of tectal development in Gymnopismultiplicata (viviparous). A,C,E: 30-mm animal. B,D,F:54-mm animal. Transverse sections of the rostral (A,B),central (C,D), and caudal (E,F) tectum. In the 30-mmanimal, cell proliferation (arrows) occurs in all parts ofthe tectum, but only the caudal tectum shows the early

pattern of cell proliferation (cell proliferation all overthe ependymal layer; arrows). In the central (C) tectum,proliferation is restricted to a lateral zone. In the rostraltectum (A), cell proliferation is highly reduced. In the54-mm animal, cell proliferation only occurs in the cau-dal tectum (arrows). Bar 5 100 µm.


are concentrated in the lateral tectum. Moremigrated cells are found in the thalamusand tegmentum.

DISCUSSION

Our examination shows that the caeciliantectum develops along the same rostrocau-dal and lateromedial gradients (Straznickyand Gaze, ’72; Schmidt and Roth, ’93) asthose of frogs and salamanders (Fig. 6). How-

ever, there are differences in the time courseof development regarding cellular prolifera-tion and the degree of cellular migration.Differences in the time course of tectal devel-opment exist not only between frogs andsalamanders, but also among caecilian spe-cies. Our results suggest that interspecificdifferences in tectal development are relatedto differences in life histories among caecil-ian taxa.

Fig. 4. A,C,E: 46-mm Typhlonectes compressicaudus(viviparous). B,D,F: 38-mm Hypogeophis rostratus (di-rect development). Transverse sections of rostral (A,B),central (C,D), and caudal (E,F) tectum. In Typhlonectes,cell proliferation is very low. Hardly any proliferative

zones can be distinguished, except in the caudal tectum(arrows). In Hypogeophis, tectal development already isnearly completed in the 38-mm animal. At this stage,there are no distinctive zones of cell proliferation. Bar 5100 µm.


Is the simplification of the tectum incaecilians due to heterochrony?

It is well known that heterochrony playsan important role in creating phylogeneticchange (Wake, ’66; Gould, ’77; Alberch, ’80,’82). Comparative studies of frogs and sala-manders provide evidence that the simpli-fied brain morphology in salamanders repre-sents the retention of early ontogenetic

features into the adult stage, thus, the adultmorphology is paedomorphic. Our resultssuggest that the simplification of tectal mor-phology in caecilians is also due to paedomor-phosis. However, tectal development in cae-cilians is not as strongly affected bypaedomorphosis as is tectal development insalamanders.

In caecilians, cellular proliferation occursall over the tectum at early ontogeneticstages but concentrates in the caudal tectumat late stages. In frogs, the homogeneouslayer of extensive cell proliferation splits offduring development, so that at later stagescell proliferation occurs in distinct medialand lateral proliferative zones (Figs. 1, 6).These two zones differ in degree of cellularproliferation. The lateral zone exhibits mostextensive cell proliferation at early ontoge-netic stages but, in this zone, cell prolifera-tion also decreases earlier than in the me-dial zone. In the medial zone, the peak ofproliferative activity is found at later devel-opmental stages and the decrease in cellproliferation occurs later than in the lateralzone. This sequence of tectal development isconcordant with that described in other ver-tebrates, such as trout (Richter, ’81; Man-sour-Robaey and Pinganaud, ’90). In sala-manders, simplification of tectal morphologyis due to a decrease in cell proliferation bothin the lateral zone and in the medial zone(Schmidt and Roth, ’93). Cell proliferation insalamanders is suppressed during late aswell as during early developmental stages(Fig. 6). Caecilians differ from frogs but aresimilar to salamanders in that, in general,the medial zone shows very low proliferativeactivity that never increases (Fig. 6), but itdecreases early in development. Consequently,the late-occurring mode of cell proliferation isdelayed in caecilians. By contrast, the patternof early cellular proliferation is similar to thatof frogs but differs from that of salamanders. Incaecilians, extensive cell proliferation withinthe lateral zone occurs early and persistsuntil mid-developmental stages (Fig. 6). Thisis in contrast to salamanders, in which cellproliferation in the lateral zone is greatlyreduced. Differences in patterns of cellularproliferation between salamanders and cae-cilians are concordant with differences inthe distribution of migrated cells at the adultstage. In the adult salamander tectum, thereare very few migrated cells in the medialand in the lateral tectum. In contrast to

Fig. 5. 66-mm Ichthyophis kohtaoensis (biphasic de-velopment). Transverse sections of the rostral (A), cen-tral (B), and caudal (C) tectum. Tectal development iswell advanced. Distinctive zones of proliferative activityhave disappeared. Bar 5 100 µm.


salamanders, migrated cells in the caeciliantectum, although few in number, are concen-trated mainly in the lateral tectum (Schmidtand Wake, ’91).

In summary, our data suggest that simpli-fication of morphological complexity in theoptic tectum of caecilians is due to a retarda-tion or loss of late developmental stages.However, in contrast to salamanders, whichshow retardation of tectal development at

early as well as at late stages, only latestages are retarded in caecilians. Thus, cae-cilians are not as strongly affected by paedo-morphosis as are salamanders.

Relationship between heterochronicdevelopment of the tectum and differences

in life historiesHeterochrony affecting the tectum occurs

in all the three amphibian orders and among

Fig. 6. Schematic drawing of ontogenetic changes inthe mode of cell proliferation in anurans, urodeles, andcaecilians. Black regions, areas of strong cell prolifera-tion; dotted regions, areas of low cell proliferation. A:Proliferation during early ontogenetic stages. B: Prolif-eration during mid-developmental stages. C: Prolifera-tion during metamorphic stages. D: Distribution of mi-grated cells in the adult tectum of the three amphibianorders. In all three orders, cell proliferation occurs allover the ependymal layer at early stages. At mid-

developmental stages, cell proliferation is restricted tolateral and medial edges of the tectum. Compared withthe situation in frogs, cell proliferation in salamandersand caecilians is highly reduced. In salamanders, thereduction especially affects the lateral zone. In caecil-ians strong cell proliferation persists in the lateral zonebut is highly reduced in the medial zone. Differences inproliferative activity among the three amphibian orderscorrelate with differences in the distribution of migratedcells within the adult tectum (D).


species of caecilians. While the general ros-trocaudal and lateromedial sequence of tec-tal development is the same, the time coursediffers among caecilians. Our data suggestthat differences in tectal development are atleast partially correlated with differences inlife histories. However, the fact that varia-tion in tectum developmental patterns alsooccurs among caecilian species that possessthe same life history suggests that heteroch-ronies are correlated with phylogenetic his-tories as well.

In order to compare patterns of heteroch-rony, criteria for the evaluation of the stageof tectal development must be established.Our criteria include the following: (1) thedegree of cellular migration, (2) the size ofthe superficial neuropil, and (3) the morpho-logical complexity of the periventricular cel-lular layer. During development, these pa-rameters differ among the rostral, central,and caudal tectum. The rostral tectum differ-entiates earliest and the caudal tectum dif-ferentiates last. Therefore a comparison ofthe above-mentioned criteria among the ros-tral, central, and caudal tectum makes itpossible to evaluate whether development iscompleted or whether it has just begun. Awell-differentiated caudal tectum indicatesthat development is already completed. Incontrast, an undifferentiated rostral tectumrepresents an early developmental stage.

Marked differences in tectal developmentoccur between the direct-developing caecil-ian Hypogeophis and all other taxa exam-ined. Our results show that in a 38-mmprehatching (15% of average adult length)Hypogeophis embryo, tectal development isfar advanced. Cell migration has occurredall over the tectum with a concentration inthe lateral part of tectum. In particular, thelate-developing caudal tectum possessesmany migrated cells. Also, a superficial neu-ropil has already developed in this region.Tectal morphology in a 38-mm Hypogeophisembryo resembles that found in a newlyhatched larva of Ichthyophis at a length of66 mm (25% of total average adult length).Thus, these species exhibit a dissociation ofdevelopment of body size and tectal develop-ment. In the viviparous caecilians Gymnopismultiplicata and Dermophis mexicanus, adevelopmental stage of tectal developmentsimilar to the 38-mm Hypogeophis is notachieved until the animal reaches lengths ofabout 84 mm and 79 mm, respectively. Thedissociation between body growth and tectaldevelopment in Hypogeophis suggests an on-togenetic repatterning that may be related

to direct development. This finding is concor-dant with the situation in salamanderswhere characteristics in the morphology ofthe feeding apparatus and the visual systemin members of the tribe Bolitoglossini (Fam-ily Plethodontidae) are attributed to an onto-genetic repatterning related to direct devel-opment (Wake and Roth, ’89). Studies onPleurodeles waltl, a salamandrid with anaquatic larval stage, demonstrate that anontogenetic repatterning similar to that ofdirect-developing bolitoglossine salamanderscan be induced by increasing thyroxine dur-ing early developmental stages (Schmidt andBoger, ’93). Changes due to thyroxine in-clude an acceleration of the rostro-caudalsequence of tectal development. The relation-ship among differences in life history, differ-ences in the action of hormones, and neuro-nal development needs investigation incaecilians. Caecilians provide an unusualparadigm for examination of the associationof hormones and life history strategies.Oviparous caecilians with free-living larvaetypically have long larval periods (e.g., Ich-thyophis glutinosus [Breckenridge and Jay-asinghe, ’79; Breckenridge et al., ’87], butperhaps not Sylvacaecilia grandisonae [Lar-gen et al., ’72]). they thus have a lengthyperiod of development dependent on endog-enous hormones. Direct-developing animalspresumably have a different interaction ofmaternal and embryonic hormones, as sug-gested by studies in frogs (Jennings andHanken, ’94) and salamanders (Schmidt andBoger, ’93). The protracted gestation periodin viviparous caecilians, with its dissocia-tion of many developmental, especially‘‘metamorphic’’ events, is probably the resultof a complex interaction of fetal and mater-nal hormones (Wake, ’89, ’94). Among thetaxa examined, Typhlonectes, a viviparousgenus, is characterized by the slowest devel-opment that especially affects the develop-ment of the neuropil and cell migration(based on a 46-mm fetus, 10% of total aver-age adult length, with fetal size represent-ing sample size, not fetal maximum). Wefind few migrated cells, and there is littleneuropil in all parts of the brain. However,relative to the development of the neuropil,the rostrocaudal retraction of proliferationzones is advanced and corresponds to thatfound in Dermophis (fetus 49 mm, 15% ofaverage adult total length) and Gymnopis(54 mm, 15% of average adult total length).There are also differences concerning thedistinction of proliferation zones. In Typhlo-nectes, no distinctive zones of cellular prolif-


eration can be distinguished except in thecaudal tectum. Differences among vivipa-rous species might be due to phylogeneticrelationships. Closely related genera, suchas Gymnopis and Dermophis, have the great-est similarities concerning the time courseof tectal development. There are very fewdifferences in the relationship betweengrowth rate (in terms of body size) and tectaldevelopment in these genera.

How do changes in the time course of tectaldevelopment in caecilians affect adult

morphology?A comparison of differences in the time

course of tectal development with the degreeof adult morphological complexity shows thatdifferent taxa may have similar adult tectalmorphologies but may vary in the timecourse of development (e.g., Ichthyophis,Typhlonectes, and Hypogeophis). Adults ofthese species possess a moderate number ofmigrated cells and a moderate degree ofmorphological complexity within the periven-tricular gray (Schmidt and Wake, ’97). De-spite these similarities, the time course ofdevelopment varies. In Hypogeophis, tectaldevelopment occurs quickly during early de-velopmental stages, whereas tectal develop-ment proceeds slowly in Typhlonectes. Themost parsimonious prediction is that an ani-mal characterized by rapid developmentshould have the greatest degree of morpho-logical complexity as an adult, and an ani-mal that develops slowly would have a lesserdegree of morphological complexity as anadult. However, similarities in the degree ofmorphological complexity between Typhlo-nectes and Hypogeophis suggest a more com-plex scenario. The degree of morphologicalcomplexity not only depends on the rate ofdevelopment, but also on the end-point ofdevelopment or the onset of a slowdown indevelopment. If the latter occurs early, theeffects of a previous acceleration might becompensated. This may be the case in Hypo-geophis, and it may be related to direct devel-opment. Similar changes in developmentalrates were demonstrated in the salamanderPleurodeles waltl after hormonal treatments(Schmidt and Boger, ’93). These studies showthat an increase in thyroxine during earlydevelopmental stages leads to an accelera-tion of development. However, due to a short-ening of the larval stage, these animals re-tain early ontogenetic characters. Insalamanders, these effects, which are in-duced by thyroxine, are considered to be

related to direct development (Schmidt andBoger, ’93). The fact that the degree of mor-phological complexity of the tectum in Typh-lonectes is similar to that in Hypogeophisindicates that despite a slower rate of devel-opment, a similar degree of morphologicalcomplexity can be achieved. Further investi-gation is needed to show whether this is dueto an extension of developmental time.

Acomparison of Hypogeophis with Gymno-pis and Dermophis shows that, despite anearly delay in development in Gymnopis andDermophis, there are more migrated cells inadults of these two genera, which may bedue to their extensive cell proliferation. Thesuperficial neuropil increases concomitantly.Even though these processes occur rela-tively later during ontogeny, in comparisonto Hypogeophis, they overcompensate for theearly delay in development in Gymnopis andDermophis.

However, in contrast to Gymnopis and Der-mophis, cell proliferation in the tectum ofthe 46-mm Typhlonectes, another viviparousspecies in a different family, is limited andthe superficial neuropil is relatively small.Typhlonectes possesses a moderate numberof migrated cells during adulthood. Suchdifferences illustrate the correlation withphylogenetic history, which may be dis-tinctly different with regard to developmen-tal patterns in taxa with the same reproduc-tive modes.

In summary, our data show that phyloge-netic history correlates with similarities inthe time course of tectal development andthus leads to a similar morphological com-plexity in the adult brain. However, similari-ties in morphological complexity do not nec-essarily predict a similar time course ofdevelopment. Alternatively, a similar degreeof morphological complexity can be achievedas a result of variation in the onset andcourse of developmental processes (cell pro-liferation, development of the neuropil). Thecritical period is presumably the end of devel-opment (larval, direct-developing embry-onic, or fetal) or the time when a generalslowdown occurs. Our data suggest that het-erochrony in caecilians is related to varia-tion in life history strategies. Further inves-tigation is needed to clarify the significanceof variation in life histories and differencesin the hormonal control involved in the regu-lation of development in amphibians.


ACKNOWLEDGMENTS

We gratefully acknowledge the technicalassistance of M. Ahlbrecht and G. Kruger.We thank W. Himstedt for Ichthyophiskohtaoensis specimens, and M. Delsol andJ.-M. Exbrayat for loan of Typhlonectes com-pressicaudus sections. We appreciate permis-sion to section heads of Gymnopis from J.Savage, and Hypogeophis from A. Grandi-son. We thank G. Roth and D.B. Wake forconstructive comments on an earlier draftof the manuscript. This work was supportedby the Deutsche Forschungsgemeinschaft(Schm 833/3-2) and the National ScienceFoundation (to M.H.W.).

LITERATURE CITED

Alberch, P. (1980) Ontogenesis and morphological diver-sification. Am. Zool. 20:653–667.

Alberch, P. (1982) The generative and regulatory roles ofdevelopment in evolution. In D. Mossakowski and G.Roth (eds.): Environmental Adaptation and Evolution.Stuttgart: G. Fischer, pp. 19–36.

Breckenridge, W.R., and S. Jayasinghe (1979) Observa-tions on eggs and larvae of Ichthyophis glutinosus.Ceylon J. Sci. 13:187–202.

Breckenridge, W.R., S. Nathanael, and L. Periera (1987)Some aspects of the biology and development of Ich-thyophis glutinosus (Amphibia: Gymnophiona). J. Zool.Lond. 211:437–449.

Cavalier Smith, T. (1978) Nuclear volume control bynucleoskeletal DNA, selection for cell volume and cellgrowth rate, and the solution of the C-value paradox.J. Cell Sci. 34:247–278.

Duellman, W.E., and L. Trueb (1986) Biology of Amphib-ians. New York: McGraw-Hill.

Fritzsch, B. (1980) Retinal projections in European Sala-mandridae. Cell Tissue Res. 225:325–341.

Fritzsch, B., W. Himstedt, and D.M. Crapon de Caprona(1985) Visual projections in larval Ichthyophiskohtaoensis (Amphibia: Gymnophiona). Dev. Brain Res.23:201–210.

Gallien, L., and M. Durocher (1957) Table chronologiquedu developpement chez Pleurodeles waltl. Michah Bull.Biol. 91:97–114.

Gilbert, S.F. (1991) Developmental Biology. Sunderland,MA: Sinauer Associates.

Gosner, K.L. (1960) A simplified table for staging anuranembryos and larvae with notes on identification. Her-petologica 16:183–190.

Gould, S.J. (1977) Ontogeny and Phylogeny. Cambridge,MA: Belknap–Harvard University Press.

Hanken, J., and B.K. Hall (1988) Skull developmentduring anuran metamorphosis: Role of thyroid hor-mone in osteogenesis. Anat. Embryol. 178:219–227.

Hanken, J., and C.H. Summers (1988a) Skull develop-ment during anuran metamorphosis: Role of thyroidhormone in chondrogenesis. J. Exp. Zool. 246:156–170.

Hanken, J., and C.H. Summers (1988b) Developmentalbasis of evolutionary success: cranial ontogeny in adirect developing anuran. Am. Zool. 28:12A.

Hedges, S.B., R.A. Nussbaum, and L.R. Maxson (1993)Caecilian phylogeny and biogeography inferred frommitochondrial DNA sequences of the 12SrRNA and16S rRNA genes (Amphibia: Gymnophiona). Herpet.Monog. 7:64–76.

Himstedt, W., and G. Manteuffel (1985) Retinal projec-tions in the caecilian Ichthyophis kohtaoensis (Am-phibia, Gymnophiona). Cell Tissue Res. 239:689–692.

Horner, H.A., and H. MacGregor (1983) C-value and cellvolume: Their significance in the evolution and devel-opment of amphibians. J. Cell Sci. 63:135–146.

Humason, G. (1979) Animal Tissue Techniques. SanFrancisco: W.H. Freeman Co.

Jennings, D.H., and J. Hanken (1994) Thyroid hormonemediation of embryonic development in a direct-developing frog, Eleutherodactylus coqui. Am. Zool.34:25A.

Largen, M.J., P.A. Morris, and D.W. Yalden (1972) Obser-vations on the caecilian Geotrypetes grandisonae Tay-lor (Amphibia Gymnophiona) from Ethiopia. Monit.Zool. Ital. 8:185–205.

Linke, R., G. Roth, and B. Rottluff (1986) Comparativestudies of the eye morphology in lungless salamanders,family Plethodontidae, and the effect of miniaturiza-tion. J. Morphol. 189:131–143.

Mansour-Robaey, S., and G. Pinganaud (1990) Quantita-tive and morphological study of cell proliferation dur-ing morphogenesis in the trout visual system. J. Hirn-forsch. 31:495–504.

Nussbaum, R.A., and M. Wilkinson (1989) On the classi-fication and phylogeny of caecilians (Amphibia: Gym-nophiona), a critical review. Herpetol. Monog. 3:1–42.

Olmo, E. (1983) Nucleotype and cell size in vertebrates:a review. Basic Appl. Histochem. 27:227–256.

Potter, H.D. (1969) Structural characteristics of cell andfiber populations in the optic tectum of the frog (Ranacatesbeiana). J. Comp. Neurol. 136:203–232.

Rettig, G., and G. Roth (1986) Retinofugal projections insalamanders of the family Plethodontidae. Cell TissueRes. 243:385–396.

Richter, W. (1981) Autoradiographische Untersuchun-gen der postnatalen Proliferationsaktivitat der Matrix-zonen des Gehirns der Forelle (Salmo irideus). Z.Mikrosk. Anat. Forsch. (Lpz.) 95:499–520.

Roth, G. (1987) Visual Behavior in Salamanders. Berlin:Springer-Verlag.

Roth, G., W. Grunwald, R. Linke, G. Rettig, and B.Rottluff (1983) Evolutionary patterns in the visualsystem of lungless salamanders (Fam. Plethodonti-dae). Arch. Biol. Med. Exp. 16:329–341.

Roth, G., C. Naujoks-Manteuffel, and W. Grunwald(1990) Cytoarchitecture of the tectum mesencephali insalamanders: a Golgi and HRP study. J. Comp. Neu-rol. 291:27–42.

Roth, G., B. Rottluff, W. Grunwald, J. Hanken, and R.Linke (1990) Miniaturization in plethodontid sala-manders (Caudata: Plethodontidae) and its conse-quences for the brain and the visual system. Biol. J.Linn. Soc. 40:165–190.

Roth, G., K.C. Nishikawa, C. Naujoks-Manteuffel, A.Schmidt, and D.B. Wake (1993) Paedomorphosis andsimplification in the nervous system of salamanders.Brain Behav. Evol. 42:137–170.

Roth, G., J. Blanke, and D.B. Wake (1994) Cell sizepredicts morphological complexity in the brains offrogs and salamanders. Proc. Natl. Acad. Sci. USA91:4796–4800.

Schmidt, A., and T. Boger (1993) Thyroxine and itssignificance for heterochronic processes in the develop-ment of the nervous system of amphibians. Ann. Anat.176:361–377.

Schmidt, A., and G. Roth (1993) Patterns of cellularproliferation and migration in the developing tectummesencephali of the frog Rana temporaria and thesalamander Pleurodeles waltl. Cell Tissue Res. 272:273–287.


Schmidt,A., and M.H. Wake (1991) Phylogenetic changesin the morphological differentiation of the caeciliantectum mesencephali. In N. Elsner and H. Penzlin(eds): Synapse—Transmission—Modulation. Proceed-ings of the Nineteenth Gottingen Neurobiology Confer-ence Stuttgart: Thieme, p. 335.

Schmidt, A., and M.H. Wake (1997) Cellular migrationand morphological complexity in the caecilian brain.J. Morphol. 231:11–27.

Schmidt, A., C. Naujoks-Manteuffel, and G. Roth (1988)Olfactory and vomeronasal projections and the path-ways of the nervus terminalis in ten species of sala-manders: A whole mount study employing the horse-radish peroxidase technique. Cell Tissue Res. 251:45–50.

Sessions, S.K., and A. Larson (1987) Developmentalcorrelates of genome size in plethodontid salamandersand their implications for genome evolution. Evolu-tion 41:1239–1251.

Straznicky, K., and R.M. Gaze (1972) The developmentof the tectum in Xenopus laevis: An autoradiographicstudy. J. Embryol. Exp. Morphol. 28:87–115.

Wake, D.B. (1966) Comparative osteology and evolutionof the lungless salamanders, family Plethodontidae.Mem. So. Calif. Acad. Sci. 4:1–111.

Wake, D.B., and G. Roth (1989) The linkage betweenontogeny and phylogeny in the evolution of complexsystems. In D.B. Wake and G. Roth (eds): ComplexOrganismal Functions: Integration and Evolution inVertebrates. Chichester: John Wiley & Sons, pp. 361–377.

Wake, M.H. (1985) The comparative morphology andevolution of the eyes of caecilians (Amphibia, Gymno-phiona). Zoomorphology 105:277–295.

Wake, M.H. (1989) Phylogenesis of direct developmentand viviparity in vertebrates. In D.B. Wake and G.Roth (eds.): Complex Organismal Functions: Integra-tion and Evolution in Vertebrates. Chichester: JohnWiley & Sons, pp. 235–250.

Wake, M.H. (1994) The concept of metamorphosis inviviparous vs. oviparous amphibians. Am. Zool. 34:93A.


Documents

Development of the tectum in Gymnophiones, with …ib.berkeley.edu/labs/mwake/papers/144.pdfDevelopment of the Tectum in Gymnophiones, With Comparison to OtherAmphibians ANDREA SCHMIDT1