Development 104, 549-556 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

549

MiJIIerian inhibiting substance production and testicular migration and

descent in the pouch young of a marsupial

JOHN. M. HUTSON1, GEOFFREY SHAW2, WAI SUM O4, ROGER V. SHORT2'3

and MARILYN B. RENFREE2

1 Department of Paediatrics, Royal Children's Hospital, University of Melbourne, Parkville, Victoria 3052, AustraliaDepartments of2Anatomy and ^Physiology, Monash University, Melbourne, Victoria 3168, Australia4Department of Anatomy, University of Hong Kong, Li Shu Fan Building, 5 Sassoon Road, Hong Kong

Summary

The ontogeny of Miillerian inhibiting substance (MIS)production by the developing testis of an Australianmarsupial, the tammar wallaby (Macropus eugenii),was determined during pouch life using an organ-culture bioassay of mouse fetal urogenital ridge. Thisinformation was related to the morphological eventsduring testicular migration and descent. MIS biologi-cal activity was found in testes (but not ovaries orliver) of pouch young from 2 to 85 days of age. MISproduction had commenced by day 2, which is withina day of the first gross morphological signs of testicu-lar differentiation. Miillerian duct regression oc-curred between 10 and 30 days, which partly coin-cided with testicular migration to the inguinal regionand enlargement of the gubernacular bulb (15 to 30days). These observations are consistent with the

hypothesis that MIS may be involved in testiculartransabdominal migration. The epididymis com-menced development and growth only after the testishad descended through the inguinal ring. This pro-vides no support for the suggestion that the epididymisis involved in testicular descent into the scrotum. Thebasic sequence of events in post-testicular sexualdifferentiation in the wallaby is sufficiently similar tothat seen in eutherian mammals to make it an excellentexperimental model for future studies of testiculardifferentiation, migration and descent.

Key words: Miillerian inhibitory substance, testiculardescent, tammar wallaby (Macropus eugenii), marsupialmammal, pouch young.

Introduction

Miillerian inhibitory substance (MIS) in eutherianmammals, including the rat, calf and human (Dona-hoe et al. 1982) causes regression of the Miillerian (orparamesonephric) duct, the anlage of the Fallopiantube, uterus, cervix and upper vagina (Josso &Picard, 1986). The hormone has been purified fromfetal (Picard & Josso, 1984) and neonatal (Budzik etal. 1985) calf testes, and is a glycoprotein of about140xl03Mr. Recently the bovine and human genesfor MIS have been isolated (Cate et al. 1986). Apartfrom Miillerian duct regression, there are no otherproven functions for MIS, although there are somedata that suggest it causes inhibition of meiosis in theovary (Takahashi et al. 1986) and may initiate testicu-lar differentiation (Vigier et al. 1987) and transab-

dominal migration of the testis from the posteriorabdominal wall to the internal inguinal ring (Hutson& Donahoe, 1986). Studies of testicular migrationand descent in the pig (Wensing & Colenbrander,1986), dog (Baumans et al. 1983) & mouse (Hutson,1986) suggest that a nonandrogenic factor, such asMIS, may control this early phase of testicular mi-gration. Subsequent descent of the testis down theinguinal canal into the scrotum appears to be underandrogenic (testosterone) control, since it usuallyfails to occur in humans (Hutson, 1986), rats (Bardin& Catterall, 1981), mice (Hutson, 1985), raccoon,dogs (Fentener van Vlissingen et al. 1984), sheep(Bruere et al. 1969) & cattle (Nes, 1966) with com-plete androgen resistance.

The mechanics of testicular descent remain poorlyunderstood, although in recent years the importance

550 /. M. Hutson and others

of the gubernaculum has been appreciated. Thegubernaculum is a cord of mesenchyme that connectsthe inguinal region to the primitive gonad and meso-nephros. With sexual differentiation and descent ofthe testis, the male gubernaculum undergoes specificchanges that are absent in the female (Backhouse,1964). During transabdominal migration of the testis,the caudal end of the gubernaculum initially enlarges,only to regress again during testicular descent into thescrotum (Wensing, 1973). This initial gubernacularswelling may be controlled by MIS, since it is abol-ished in male mice fetuses exposed to diethylstilbo-estrol (DES), which also prevents regression of theirMiillerian ducts (Raynaud, 1958).

Tammar wallabies (Macropus eugenii) weigh only440 mg on the day of birth. At this stage, the gonadsof males and females are still attached to the func-tional mesonephric kidney high in the abdomen(Tyndale-Biscoe & Renfree, 1987) and are not mor-phologically different by light microscopy and mor-phometry (O et al. 1988). Testicular morphologychanges rapidly after birth and seminiferous tubulesare clearly seen in the gonads of males by day 2postpartum (Short et al. 1988). Thus while the youngare easily accessible within the pouch, testiculardifferentiation, migration and descent occur (Tyn-dale-Biscoe & Renfree, 1987). Since the timing ofthese events in marsupials differs from the commoneutherian pattern, the aims of this study were toprovide preliminary data on the timing of MISproduction by the testes, and to relate this to thedevelopment of the testis and gubernaculum andtesticular migration and descent.

Materials and methods

Animals62 pouch young of known age and sex determined byphenotypic appearance and/or karyotype using the methoddescribed by O et al. (1988) were obtained from thebreeding colony of tammar wallabies (Macropus eugenii)maintained at Monash University. The day of birth isdesignated as day 0.

MorphologyThe young tammars were removed from the pouch andtransferred in an insulated warm container to the labora-tory, where they were killed by decapitation. The abdomi-nal cavity was opened under sterile conditions and the rightgonad removed for MIS bioassay. Care was taken topreserve the normal anatomy on the left side of the animal.The caudal half of the pouch young was fixed in Bouin'ssolution for at least one week and stored in 70% ethanol.After trimming the specimens, they were embedded inparaffin blocks and serial coronal or transverse sectionswere cut at 6/im and stained with haematoxylin and eosinor Masson's trichrome.

Sections were examined under a dissecting microscopeand drawn with the aid of a camera lucida or by micropro-jector, at magnifications ranging from x30 to xlOO.

BioassayGonads and liver from pouch young of different ages (2-91days old) were weighed and then incubated in an organ-culture system to detect MIS biological activity (Donahoe etal. 1977). The standard MIS bioassay was modified to usethe 13i-day fetal mouse (instead of rat) urogenital ridge,which was cocultured with the wallaby tissue for 72 h on anagar-coated grid over 0-7 ml of CMRL medium (Gibco,NY, USA) with 100 i.u. penicillin and 200 jtg streptomycin(Gibco, NY, USA) and 10 % fetal calf serum (Gibco, NY,USA). Cultures were maintained at 37°C in a humidifiedatmosphere of 5 % CO2 and 95 % air, following which thetissue was embedded in 2 % agar. The specimens were thenfixed in Bouin's solution, dehydrated in alcohol, cleared inxylene and embedded in paraffin. Serial sections (8jim)were stained with haematoxylin and eosin. The degree ofregression of each mouse Miillerian duct was scored subjec-tively from zero (no regression) to 5 (complete regression)by two independent observers after examination of sectionsspanning the entire duct length. In early regression (grade1), the duct is smaller than normal and the basementmembrane has begun to dissolve. The mesenchyme aroundthe duct forms a loose whorl. With increasing regression,the diameter of the duct shrinks further to about half thenormal size (grade 2). The lumen then decreases in size(grade 3) or becomes obliterated (grade 4), leading to totalregression (grade 5).

MIS antibody preparationAntiserum to MIS was raised in three New Zealand whiterabbits against purified bovine MIS donated by Dr P. K.Donahoe (Massachusetts General Hospital, Boston). Theinitial immunization was performed with Freund's com-plete adjuvant and subsequent boosts were given withFreund's incomplete adjuvant. All injections were subcu-taneous. Immunoglobulin-G was isolated from the com-bined MIS antiserum by incubation with protein-A-sepha-rose (Sigma) followed by elution with citrate buffer ( 0 1 M ,pH3-5). This igG fraction was dialysed against 10mM-phosphate buffer containing 0-15M-NaCl (PBS, pH7-4)and SDS-PAGE was performed to check the purity of thetotal IgG fraction. Either 25 /.d or 250/il of concentratedIgG was added to the organ-culture medium, followingpreliminary studies showing that 250 1̂ was able to inhibitMIS activity from bovine and mouse testis.

Results

MIS production

MIS biological activity was found in almost all testesexamined, which were from pouch young 2 to 85 daysold (Fig. 1). In contrast, the ovaries (3-91 days) andpieces of liver from males or females contained nomeasurable MIS (results not shown). Rabbit poly-clonal antibodies raised against bovine MIS were

Mouse

•o03C

10r

86420

11-28 days

n.15

in

5

n liiii::

•21-40 days

it I i i i a

41-91 days

0 1 2 3 4 5MIS bioactivity

Fig. 1. MIS bioactivity of wallaby pouch young gonadsfrom 2 to 91 days after birth. The numbers of gonadsfrom male pouch young (stipple bars) and female pouchyoung (open bars) with a given grade of MIS bioactivityare shown, in a series of age groups. MIS activity of themouse testes used as assay controls is also shown (topgraph, stipple). Liver specimens had no measurable MISactivity and are not shown.

tested for their ability to inhibit the biological activityof wallaby MIS after concentration of the IgG frac-tion of the antiserum to 25-3 mgml"1. When 25 iA ofthe concentrated IgG was added to the bioassaymedium (0-7ml), no significant inhibition of MISactivity was found. When 250 /zl was added, however,MIS activity was abolished in seven out of eightassays (Fig. 2) (P<0-01 , Wilcoxon rank-sum test).

Testicular migration and descentIn the day 2 postpartum male tammar, the testis wasrecognizable as such by the formation of primitiveseminiferous cords. The testis and mesonephros werelocated within the abdominal cavity, with the 0-5 mmdiameter testis on the medial side of the larger

MIS production in a marsupial 551

(0-75x2-5 mm) mesonephros. The lower pole of thetestis and the caudal end of the mesonephros wereattached to the inguinal region by a 0-1 mm thickgubernacular cord, 1 mm long (Fig. 3A). Two scrotalbulges were easily identified on the ventral body wall,anterior to the pubis. The gubernaculum extendedthrough the developing body wall into the scrotalbulges, with the apex of the processus vaginalis withinthe gubernaculum extending about halfway from theinguinal canal to the scrotum. Distally, the processuswas a narrow slit, but near the external obliquemuscle the space was wider and filled with thegubernaculum proper.

By day 10 after birth, the testis had well-formedseminiferous tubules and it had increased to aboutlmm in length while the mesonephros remainedunchanged in size. Although the caudal end of theMiillerian duct had not reached the urogenital sinusby day 10, the cranial root showed early signs ofregression, including decrease in size, breakdown ofthe basement membrane and a loose whorl of sur-rounding mesenchyme. The intra-abdominal cord ofthe gubernaculum was still 1 mm long (Figs 3B, 5A),although the extra-abdominal gubernaculum ex-tended into the fused scrotal bulges. The apex of theprocessus vaginalis invaded the centre of the guberna-culum down to the neck of the scrotum (Fig. 4A).The gubernaculum proper within the processus hadenlarged to twice its diameter at 2 days, forming anextra-abdominal 'bulb'. The peripheral (vaginal) partof the gubernaculum contained myoblasts differen-tiating into the cremaster muscle.

By day 15 to 17, the testis had elongated to1-1-5mm, but its lower pole remained about lmmfrom the inguinal canal (Fig. 3C). The excretorytubules of the mesonephros had begun to involuteand the mesonephros now was a similar size to that ofthe testis (0-5x1-5 mm). The Miillerian ducts were inadvanced regression. The two scrotal swellings werecompletely fused and now formed a bulbous projec-tion anterior to the pubis. The gubernacular mesen-chyme was enlarging, with the extra-abdominal bulbnow about four times bigger (0-25x0-6mm) than atday 2 (Fig. 4B).

By day 29, the testis (0-75x1-25mm) had movedfrom its intra-abdominal position through the ingui-nal canal into the extra-abdominal space above theneck of the scrotum (Fig. 3D), which was a well-formed structure lmm in diameter. The excretorytubules of the mesonephros had involuted com-pletely, leaving the mesonephric ducts, which had notyet enlarged by elongation and convolution to formthe epididymis. The Miillerian duct had almost com-pletely regressed. The processus vaginalis extendedinto the midscrotal gubernacular mesenchyme, whilethe bulb of the gubernaculum had enlarged still

552 J. M. Hutson and others

Fig. 2. Mouse urogenital cords cocultured with wallaby gonads. (A) An ovary from a 9-day-old pouch young induces noregression of the Mullerian duct (Md) (grade 0 regression). The Wolffian duct has regressed in the absence of androgensupport. (B) A testis from a 19-day-old pouch young causes grade 2-3+ Mullerian duct regression, and also stimulatesthe Wolffian duct (Wd). (C) In the presence of 25 fi\ of rabbit anti-bovine-MIS IgG concentrate a testis from a 19-day-old pouch young causes Miillenan duct regression grade 2+. (D) 250 f.A of IgG is sufficient to block the MIS activity of atestis from a 19-day-old pouch young (grade 0 Mullerian duct regression). Slides are stained with haematoxylin andeosin. Bar, 0-1 mm.

further to O-3xO-6mm (Figs 4C, 5B).The testis had reached the neck of the scrotum by

day 44. By this time, the scrotum had grown to 3 mmin diameter, the apex of the processus reached to thebottom of it, and the cremaster muscle was welldeveloped. The bulb of the gubernaculum had begunto shrink, now measuring 0-4 mm in diameter(Fig. 4D), with its total length from scrotal attach-ment to the developing epididymis being 1-5 mm.

Descent of the testis was essentially complete byday 64-68, when it was located near the bottom of thescrotum. Further testicular growth had occurred,since it now measured lxl-5mm. The caput of theepididymis was enlarging rapidly by elongation of themesonephric ducts. The vaginal layer of the guberna-culum had condensed around the processus vaginalis,especially cranial to the testis, to form a tight fascialsheath for the spermatic cord. The diameter of the

MIS production in a marsupial 553

Fig. 3. Position of the testis at various ages in cameralucida drawings from coronal sections. (A) In the 2-day-old male, the testis (/) is 1 mm cranial to the internal ringof the inguinal canal, the position of which is marked bya broken line. The gubernaculum (g) passes through theinternal inguinal ring. The testis is much smaller than themesonephros (m) which is still functional. The bladder(b) and pubic bones (pb) are also indicated. (B) By day10, the testis has enlarged substantially, but still remainsabout 1 mm cranial to the internal inguinal ring. TheMullerian duct (Md) and Wolffian duct (Wd) are bothwell developed at this stage (also see thephotomicrograph in Fig. 5A). (C) At day 17, the testis isstill about 1 mm from the internal inguinal ring. Themesonephros is regressing, but the mesonephros andtestis are still too large to pass together through theinguinal ring. The Mullerian duct is regressing, whilst theWolffian duct is well developed. The processus vaginalis(pv) can be seen at the inguinal ring in this section.(D) By day 29, the testis lies within the inguinal canal.The Miillerian duct is almost totally regressed.

processus was less than at the time of testicularmigration. The gubernaculum was shorter than pre-viously observed and was oriented differently, thetestis now hanging down from the involuting guberna-culum rather than being cranial to it.

Fig. 4. Development of the extra-abdominalgubernaculum in male pouch young (camera-lucidadrawings from coronal sections). (A) Day 10. Thegubernaculum (g) extends to the areolar tissue (at) in thescrotal bulges (scb). The caudal end of the processusvaginalis (pv) is seen at the base of the scrotal bulges.(B) Day 17. The processus vaginalis develops into anannular slit in the inguinal region with the growth of thebulb of the gubernaculum. Subcutaneous areolar tissue(of), abdominal muscles (am) and pubis (p) are alsoshown. (C) Day 29. The gubernacular bulb is furtherenlarged, accompanied by growth of the scrotum (seephotomicrograph in Fig. 5B). (scs, scrotal skin). (D) Day47. The gubernacular bulb has ceased growth, despitecontinued enlargement of the scrotum.

Discussion

The developing testes of tammar pouch young pro-duce MIS; tammar testes caused significant Mullerianduct regression in the mouse bioassay and rabbitpolyclonal antibodies raised against purified bovineMIS blocked this biological activity. No MIS bioacti-vity was found in ovaries of pouch young aged 3-91days; antral follicular development does not begin inthe ovaries until day 110 of pouch life (Tyndale-Biscoe & Renfree, 1987). These results are in accordwith previous reports that MIS from a wide variety ofmammalian and avian species can cause Mullerianduct regression in the rat or mouse embryo (Donahoeet al. 1982; Josso & Picard, 1986), and suggest that thebiochemistry and function of marsupial MIS is suf-

554 J. M. Hutson and others

P

Testis differentiation

MIS secretionMullerianregression

Transabdominalmigration

Inguinoscrotaldescent

Fig. 5. (A) Photomicrograph of the urogenital ridge of a day-10 pouch young, showing the testis (t) in association withthe mesonephros (me). The gubemaculum (g) extends through the internal inguinal ring (ir) just off the edge of themicrograph in the abdominal muscles (aw). The Wolffian duct is well developed. Bar, 0-5 mm. (B) The gubernacularbulb (gb) in the scrotal region of a day-29 male pouch young. The cremaster muscle (cm) is developing in the outer rimof the gubemaculum. pv, processus vaginalis; at, areolar tissue of scrotum. Masson's trichrome stain. Bar, 01 mm.

ficiently homologous with that of eutherian mammalsto make marsupials a useful experimental model.

The testes of wallaby pouch young produced bio-logically active MIS in all samples examined fromday 2 to day 85. The onset of MIS productiontherefore precedes Mullerian duct regression whichoccurs between days 10 and 30. This is consistent withthe observation in rats that the Mullerian duct issensitive to MIS only during a short period (Donahoeet al. 1982). The transabdominal migration of thetestis and enlargement of the gubernacular bulboccurs at a similar time, between 15 and 30 days(Fig. 6). This is consistent with the hypothesis thatMIS is involved in these functions.

Testicular androgen production is probably occur-ring over this same period, since the Wolffian ductregresses in females but not males between days 10and 25. However, in eutherians transabdominal mi-gration of the testis and gubernacular outgrowth arenot dependent on androgens (Baumans et al. 1983),leading to suggestions that they might be under the

t 5Birth

20 40Age (days)

60

Fig. 6. Diagram summarizing the temporal relationshipbetween testicular development, MIS secretion,Mullerian duct regression, and testicular migration anddescent in tammar pouch young.

control of a nonandrogenic hormone such as MIS(Baumans et al. 1983; Hutson, 1985).

MIS production has already begun by day 2 post-

MIS production in a marsupial 555

partum, the earliest age tested, when we see the firstappearance of seminiferous cords in the developingmale gonad (Short et al. 1988). This is comparable tothe fetal rat, where MIS production begins at 13 daysof gestation, which is at the commencement ofmorphological differentiation of the testis (Jost, 1972;Tran et al. 1987). If in future studies we can show thatthe onset of MIS secretion precedes morphologicaldifferentiation of the testis, this would support a rolefor MIS in testicular differentiation (Vigier et al.1987).

The ontogeny of testicular migration and descent inthe tammar wallaby is similar to that seen in pigs(Wensing & Colenbrander, 1986) and humans (Back-house, 1982). In particular, the anatomical develop-ment of the gubernaculum is similar to that in othermammalian species. There is an outgrowth phase,where the extra-abdominal (caudal) gubernaculumenlarges to form a 'bulb', while the cord-like cranialgubernaculum shortens. Wensing & Colenbrander(1986) have proposed that the gubernaculum is themediator of testicular descent and, in the tammar, thegubernacular enlargement is associated with transab-dominal migration of the testis over a distance oflmm. After the testis passes through the abdominalwall via the inguinal canal, the gubernaculum be-comes progressively smaller, which is similar to the'regression' phase described in pigs (Wensing &Colenbrander, 1986).

The development of the wallaby epididymis isdifferent from that in eutherians, particularly themouse (Hadziselimovic et al. 1978). Because themarsupial mesonephros persists for some time afterbirth as a functional kidney (Tyndale-Biscoe & Ren-free, 1987) the subsequent differentiation of its ductsystem into an epididymis is delayed, as in chickembryos, where the mesonephros functions untilhatching (Romanoff, 1960). This is completelydifferent from the mouse and human where themesonephros involutes much earlier in fetal develop-ment (Hadziselimovic & Kruslin, 1979). Hadziseli-movic has proposed that enlargement of the caput ofthe epididymis may push the testis into the scrotum,thereby being the cause of testicular descent (Hadzi-selimovic & Kruslin, 1979). This hypothesis cannothold true for the tammar wallaby, where the epididy-mis does not even begin to develop until about thetime the testis enters the scrotum. In addition, thetammar testis cannot enter the inguinal canal (whichhas a diameter of about 0-5 mm just before descent)until after the mesonephros has involuted. Inguinalpassage of the testis, therefore, appears to be moredependent on mesonephric regression, rather thanepididymal enlargement.

The tammar wallaby pouch young offers a uniqueopportunity to study gonadal differentiation and

testicular migration and descent without the need forcomplex intrauterine surgery and without the con-founding variable of an endocrine placenta. In thispaper, we have attempted to describe the normalsequence of events; further studies are now in pro-gress to investigate the role of MIS in testiculardifferentiation, migration and descent.

The authors thank Lisa Watts and Pam Farmer forexcellent technical assistance, and Dani Blanden and DrT. P. Fletcher for assistance with the wallabies. The projectwas supported in part by grants from the Royal Children'sHospital Research Foundation, the John Winter Foun-dation, University of Melbourne, the National Health andMedical Research Council, and the Australian ResearchGrants Scheme. The wallabies are held under licence no.84-28 from the Victorian Department of Conservation,Forests and Lands.

References

BACKHOUSE, K. M. (1964). The gubernaculum testisHunteri: testicular descent and maldescent. Ann. Roy.Coll. Surg. Engl. 35, 15-33.

BACKHOUSE, K. M. (1982). Embryology of testiculardescent and maldescent. Urol. Clin. Nth. Am. 9,315-325.

BARDIN, C. W. & CATTERALL, J. F. (1981). Testosterone:A major determinant of extragenital sexualdimorphism. Science 211, 1285-1294.

BAUMANS, V., DIJKSTRA, G. & WENSING, C. J. G. (1983).

The role of a non-androgenic testicular factor in theprocess of testicular descent in the dog. Int. J. Androl.6, 541-552.

BRUERE, A. N., MCDONALD, M. F. & MARSHALL, R. B.

(1969). Cytogenetical analysis of an ovine malepseudohermaphrodite and the possible role of the 7chromosome in cryptorchidism of sheep. Cytogenetics8, 148-157.

BUDZIK, G. P., DONAHOE, P. K. & HUTSON, J. M. (1985).

Purification of Mullerian inhibiting substance and amodel for its mechanism of action. Prog. Clin. Biol.Res. 171, 207-223.

CATE, R. L., MATTALIANO, R. J., HESSION, C , TIZARD,

R., FARBER, N. M., CHEUNG, A., NINFA, E. G., FREY,

A. Z., GASH, D. J., CHOW, E. P., FISHER, R. A.,

BERTONIS, J. M., TORRES, G., WALLNER, B. P.,

RAMACHANDRAN, K. L., RAGIN, R. C , MANGANARO,

T. F., MACLAUGHLIN, D. T. & DONAHOE, P. K. (1986).

Isolation of the bovine and human genes for Mullerianinhibiting substance and expression of the human genein animal cells. Cell 45, 685-698.

DONAHOE, P. K., BUDZIK, G. P., TRELSTAD, R. L.,

MUDGETT-HUNTER, M. , FULLER, JR., A . F . , HUTSON,

J. M., IKAWA, H., HAYASHI, A. & MACLAUGHLIN, D.

(1982). Milllerian inhibiting substance: an update.Recent Prog. Horm. Res. 38, 279-326.

DONAHOE, P. K., ITO, Y. & HENDREN, W. H. Ill (1987).

A graded organ culture assay for the detection of

556 J. M. Hutson and others

Mullerian inhibiting substance. J. Surg. Res. 23,141-148.

FENTENER VAN VLISSINGEN, M., COLENBRANDER, B.,

VERBRUGGEN, A. & WENSING, C. (1984). Testicularfeminized males (TFM) in Nyctereutes procyonoides(Racoon Dog). INSERM 123, 335-340.

HADZISELIMOVIC, F., HERZOG, B. & KRUSLIN, E. (1978).The morphological background of estrogen-inducedcryptorchidism in the mouse. Folia Anat. Jugos. 8,63-74.

HADZISELIMOVIC, F. & KRUSLIN, E. (1979). The role ofthe epididymis in descensus testis and the topographicalrelationship between the testis and epididymis from thesixth month of pregnancy until immediately after birth.Anat. Embryol. 155, 191-196.

HUTSON, J. M. (1985). A biphasic model for thehormonal control of testicular descent. Lancet 2,419-421.

HUTSON, J. M. (1986). Testicular feminization: a modelfor testicular descent in mice and men. J. Pediatr. Surg.21, 195-198.

HUTSON, J. M. & DONAHOE, P. K. (1986). The hormonalcontrol of testicular descent. Endocr. Rev. 7, 270-283.

Josso, N. & PICARD, J.-Y. (1986). Anti-Mullerianhormone. Physiol. Rev. 66, 1038-1090.

JOST, A. (1972). Donnees preliminaires sur les stadesinitiaux de la differenciation du testicule chez le rat.Arch. Anat. Microsc. Morphol. exp. 61, 415-438.

NES, N. (1966). Testikulaer feminisering hos Storfe.Nord. Vet.-Med. 18, 19-29.

O, W.-S., SHORT, R. V., RENFREE, M. B. & SHAW, G.(1988). Primary genetic control of somatic sexualdifferentiation in a mammal. Nature, Lond. 331,716-717.

PICARD, J.-Y. & Josso, N. (1984). Purification oftesticular anti-Miillerian hormone allowing directvisualization of the pure glycoprotein anddetermination of yield and purification factor. Molec.

Cell Endocr. 34, 23-29.RAYNAUD, A. (1958). Inhibition, sous l'effet d'une

hormone oestrogene, du developpement dugubernaculum du foetus male de souris. C.r. hebd.Seanc. Acad. ScL, Paris. 246, 176-179.

ROMANOFF, A. L. (1960). The Avian Embryo. New York:Macmillan. 781 pp.

SHORT, R. V., RENFREE, M. B. & SHAW, G. (1988).Sexual development in marsupial pouch young. InDevelopment of Young Marsupials: Models forBiomedical Research (ed. C. H. Tyndale-Biscoe &P. A. Janssen), pp. 200-210. Berlin: Springer-Verlag.

TAKAHASHI, M., KOIDE, S. S. & DONAHOE, P. K. (1986).Mullerian inhibiting substance as an oocyte meiosisinhibitor. Molec. Cell Endocr. 47, 225-234.

TRAN, D., PICARD, J. Y., CAMPARGUE, J. & Josso, N.(1987). Immunocytochemical detection of anti-Miillerian hormone in Sertoli cells of variousmammalian species including human. J. Histochem.Cytochem. 35, 733-743.

TYNDALE-BISCOE, C. H. & RENFREE, M. B. (1987).Reproductive Physiology of Marsupials. Cambridge:Cambridge University Press. 476pp.

VIGIER, B., WATRIN, R., MAGRE, S., TRAN, D. & Josso,N. (1987). Purified bovine AMH induces acharacteristic freemartin effect in fetal rat prospectiveovaries exposed to it in vitro. Development 100, 43-55.

WENSING, C. J. G. (1973). Testicular descent in somedomestic mammals. II. The nature of thegubernaculum changes during the process of testiculardescent in the pig. Proc. Kon. Ned. Akad. Wetensch.C. 76, 190-195.

WENSING, C. J. G. & COLENBRANDER, B. (1986). Normaland abnormal testicular descent. Oxford Rev. Reprod.Biol. 8, 130-164.

(Accepted 2 September 1988)