/ . Embryol. exp. Morph. Vol. 33, 7, pp. 13-28, 1975 \ 3

Printed in Great Britain

Experimental cardiac morphogenesis

I. Development of the ventricular septum in the chick

By JUNG Y. HARH1 AND MILTON H. PAUL2

From the Department of Pediatrics, Johns Hopkins University Schoolof Medicine, Baltimore, and the Department of Pediatrics, Northwestern University

Medical School, Chicago

SUMMARY

An experimental technique utilizing microfiber markers and radioautography was usedto study the morphogenesis of the ventricular septum in the chick embryo heart from stages22 to 33 of Hamburger & Hamilton (1951). Fibers placed in the myocardium of the primitiveventricles of 4-day-old embryos within 250 /tm to each side of the future site of the ventricularseptum resulted in shortening of the distance between the two fibers until both were foundwithin the myocardium of the septum at 7 days. The fibers coated with tritiated thymidinelabeled the myocytes of the trabeculae immediately adjacent to where the fibers were placedand showed that when trabeculae were labeled within the 400-500 /tm width centered on theventricular septum, they aggregated together. The labeled myocytes in the trabeculae werefound from the smooth crest to the most apical portion of the septum. These findings suggestthat the muscular ventricular septum is formed by aggregation and coaptation of trabeculaeand is of a single developmental origin.

INTRODUCTION

The ventricular septum in the chick embryo is first noted approximately atfour days of incubation as a loose meshwork of trabeculae. Between day 4 and 8of incubation, this meshwork of trabeculae solidifies into the muscle mass thatis the muscular ventricular septum in the mature heart. A portion of the heartloop is thus separated into areas that will later become the definitive right andleft ventricles.

A number of studies has been reported to resolve the origin of the ventricularseptum in both human and experimental animals. These studies representdescriptive embryology based upon information from histologic sections ofembryonic hearts. They have given rise to a number of different views on howthe septum arises (Flack, 1909; Tandler, 1912; Murray, 1919; Waterston, 1919;Takahashi, 1923; Frazer, 1932; Kramer, 1942; Streeter, 1948; Grant, 1962;

1 Author's address: University of Virginia Hospital, Department of Pediatrics, Charlottes-ville, Virginia 22903, U.S.A.

2 Author's address: Department of Pediatrics, Northwestern University Medical School,The Children's Memorial Hospital, Willis J. Potts Children's Heart Center, Chicago, 111.,U.S.A.

14 J. Y. HARH AND M. H. PAUL

De Vries & Saunders, 1962; Van Mierop, Alley, Kausel & Stranahan, 1963;Patten, 1964; Van Mierop & Netter, 1969; Goor, Edwards & Lillehei, 1970).Much of this controversy is based upon limitations in the techniques utilized.

We have been able to label specific cells of the heart by inserting microfiberscoated with tritiated thymidine in the chick embryo heart during the periodin which ventricular septation is taking place. Labeling specific locations ofthe embryo with a durable, non-diffusible marker circumvents many of thecriticisms raised against the earlier investigations.

MATERIALS AND METHODS

White Leghorn eggs were used throughout this study. Fertile eggs wereincubated at 38 ±0-5 °C for appropriate periods and all embryos were stagedaccording to the criteria of Hamburger & Hamilton (1951). After candling eachegg, a window was made in the shell and the vitelline and amniotic membraneswere torn with forceps, exposing the pericardium of the embryo heart. Theoperation on the embryonic heart was made in the manner described below.The shell opening was then sealed with parafilm and melted paraffin and theegg was returned to the incubator.

Microfiber markings. Sterile nylon fibers (black braided suture, Dektanel)approximately 12 /on in diameter were implanted into the embryonic myo-cardium perpendicular to the surface of the ventricle, using ultrafine forcepsunder a dissecting microscope (x 4-25). The distal end of each fiber was barelyintruded into the ventricular cavity; the proximal end, which protruded fromthe surface, was trimmed even with the surface so that the final length of eachfiber was approximately 200 /tm. Particular care was taken to minimize intrusionof the fibers into the ventricular cavity since it is possible for such intrusion todistort the embryonic bloodstreams (Harh et ah 1973).

A pair of fibers was placed in each embryo in one of the following three ways:group I, one fiber in each primitive ventricle to each side of the future site ofthe ventricular septum; group II, both fibers in the primitive right ventricle;group III, both fibers in the primitive left ventricle. The distance between thetwo fibers in each embryo was measured with a micrometer under directvisualization, and this measurement was repeated daily. To reduce error inmeasuring the distance between fibers, the embryos were lifted up with aninstrument to stop the pulsation of the heart for a second or so. Embryos thatsurvived were autopsied at 7 days of embryonic age and those that died less than48 h post-operatively or those in which one or both fibers had been extruded,were excluded from the final analysis. Following gross external examination of allthe embryos, they were fixed in Bodian's solution (80 % ethanol, 90parts; glacialacetic acid, 5 parts; formaldehyde, 5 parts), and final examination was madeunder a dissecting microscope.

Tritiated thymidine-coating techniques. Sterile nylon fibers approximately

Ventricular septum in chick 15

Table 1. One fiber in each ventricle

Embryo no.

VS 114VS 120VS 129VS 135VS 154VS188VS217VS219VS232VS241VS279VS280VS624VS628VS637VS649

Age at the time*of operation

4 D -4D +4D-4 D4D +4 D4D +4D +4D-4D +4 D4 D4D-4D-4D +4D-

cOh

275400450400350300425350300350300400500400450350

Distance variationsthe two fibers

K

24 h

150Lf400400300300300300150 Rf200 Lf200400300425250350250

> betweentOm)

48 h

RLf250300250150 RLf175 RLf150LfRLfRLf125 RLf300 RfRLf300175 Rf250 Rf200 Lf

72 h

RLf————————RLfRLfRLfRLfRLfRLf

* 4 D + /— indicates 4 days plus or minus 3 h.t Rf or Lf implies the fiber in the pi imitive right and left ventricle is already incorporated

into the septal myocardium. RLf without distance indicates both fibers are already withinthe septal myocardium and not visible on the surface. They were found only when the heartswere dissected under the microscope.

12 /tm in thickness were dipped in concentrated thymidine-methyl H3, 20 Ci/mmol/c.c. (New England Nuclear); dehydrated using an air dryer, then left atroom temperature overnight. The fibers were implanted into the embryonicmyocardium in the manner described above for unlabeled fibers and the embryoswere reincubated for 6 h (36 embryos), 24 h (27 embryos), 48 h (37 embryos),72 h (49 embryos), and 84 h (57 embryos) so that the resulting total incubationperiod was 4£, 5, 6, 7, and 1\ days respectively. The sacrificed embryos werefixed in Bodian's solution and prepared for serial section at 10 /im thickness.Deparaffinized slides were dipped in Kodak type 3 NTB liquid emulsion.After 30-45 days' exposure, the sections were then developed in Kodak D-19developer, fixed in Kodak fixer, stained in 0-1 % nuclear fast red and 1 %picric acid and examined for labeled silver granules under the microscope(Belanger & Leblond, 1946; Gross, Bogoroch, Nadler & Leblond, 1951;Messier & Leblond, 1957; Sissman, 1966; Rosenquist & DeHaan, 1966;Feitelberg & Gross, 1970).

RESULTS

Although it is possible for the fibers to cause mechanical distortion of theembryonic bloodstreams during the early developmental stages of the ventricular

16 J. Y. HARH AND M. H. PAUL

Fig. 1. Experimental group I embryonic chick heart. Note the change of distancebetween the two fibers when they were placed approximately 400 ftm apart (arrowsindicate the fibers). A, At the time of operation; B, 1 day after fiber placement;C, 2 days after fiber placement; D, 3 days after fiber placement (from below).RV, Primitive right ventricle; LV, Primitive left ventricle, x 18.

Ventricular septum in chick 17

A

Fig. 2. Experimental group I embryonic chick heart. Fibers located at the centerof the septum (B) were originally placed 200/tm apart. A, At the time of operation;B, 3 days after fiber placement, x 17.

Table 2. One fiber in each ventricle

Embryo no.

VS .113VS115VS187VS 190VS 192VS193VS195VS197VS223VS277VS284VS663

Age at the time*of operation

4 D -4 D -4 D4 D4 D +4 D +4 D +4 D +4 D4 D -4D +4D +

Oh

225225200200200200250250250250250225

Distance variations betweenfthe two fibers (/tm)

A

24 h

125 Rf225250150Rf100 RLf150 Rf200 Rf150 Rf175 Rf150 RLf200200 Lf

48 h

RLf150150 RLf75 RLf

RLfRLfRLfRLf100 RLfRLf

50 RLf150 RLf

72 h

RLf——————RLf—RLfRLf

* 4 D + /— indicates 4 days plus or minus 3 h.t Rf or Lf implies the fiber in the primitive right or left ventricle is already incorporated

into the septal myocardium. RLf without distance indicates both fibers are already withinthe septal myocardium and not visible on the surface. They were found only when the heartswere dissected under the microscope.

EMB 33

18 J. Y. HARH AND M. H. PAUL

Table 3. One fiber in each ventricle

Embryo no.

VS 116VS 124VS 125VS627VS633VS639VS 641VS644VS650VS662VS666VS668VS673VS674

Age at the time* rof operation

4 D4 D +4 D +4 D -4 D +4 D +4 D +4 D -4 D +4 D +4 D +4 D -4 D -4 D

* 4 D + / - indicates

Oh

600800600600650800600550600725550550625600

4 days

Distance variations betweenthe two

24 h

750900500775700

1000700675700800600650675700

plus or minus

fibers (/*m)

48 h

300550300500450900450500700500350300400400

3h.

72 h

400—300—350——350575——275300—

system, the fibers do not appear to disturb normal development of the heart inour experiments. Both the hearts in which the fibers were extruded from theventricles immediately following operation and the hearts in which the fibersremained for additional periods of incubation developed normal and normallyrelated great vessels. This is not surprising since the fiber is only 12 /m\. indiameter, which is less than the diameter of a single trabecula of the ventricleof the 4-day-old embryo.

Microfiber markings. In group I embryos (42), when the fibers were placed nofurther than 200-250 /«n to each side of the center of the future site of theventricular septum (Table 1), the distance between the two fibers was shorteron each subsequent examination until at 7 days both fibers were found withinthe ventricular septum. In those hearts in which the fibers were not visible onthe surface at 7 days of incubation, they were found within the septal myocar-dium when it was dissected under the dissecting microscope (Figs. 1 and 2). Thecloser the fibers were placed to the site of the future ventricular septum, theearlier they were embedded in the septum (Table 2). As a result, they werefound at the septum most centrally. All the fibers found within the septumpointed toward the site of the interventricular foramen. However, when thefibers were placed further than 250 ^m to each side of the site of the futureventricular septum, the distance between them increased by the following day.Subsequently the distance between them decreased, but these fibers remainedin the free wall of the ventricular myocardium and never became incorporatedinto the septum (Table 3). In some embryos in which the fibers were placed

Ventricular septum in chick 19

Fig. 3. Experimental group II embryonic chick heart (right lateral view). Notethe change of distance when the fibers were placed in the primitive right ventricle(arrows indicate the fibers). A, At the time of operation; B, 1 day after fiber place-ment; C, 2 days after fiber placement; D, 3 days after fiber placement, x 18.

20 J. Y. HARH AND M. H. PAUL

Table 4. Two fibers in the right ventricle

Embryo no.

VS119VS169VS 212VS220VS227VS230VS285VS307VS3O8VS312VS 319VS322VS323VS343VS348VS35OVS351VS354VS359VS664VS665VS669VS670VS672VS675VS676VS679VS684

Age at the time*of operation

4D +4 D -4 D4D +4D +4 D -4 D -4 D -4 D4 D -4D-4 D -4D +4D +4D-4D-4D-4D +4D +4D +4D-4D-4D-4 D4D +4D +4D +4D-

*4D + / -

Oh

500250300400400250300150200200100300250.175225200150100175300275125150150250200200125

indicates 4

Distance variations betweenthe two fibers (/*m)

24 h

800300475600500350400200250200250500300250300250225200300525350200250200400325250175

days plus or

48 h

1200450500850700500550400400450400700400450550400450350550625650450450400700500525450

minus 3 h.

72 h

—650——950750—575—625—900600—725700650800—900900750650675950—650—

within 250/tm of the center of the future ventricular septum, the distancebetween the two fibers remained unchanged or increased after 24 h of incubation,and subsequently decreased. Interestingly in these hearts, the ventricular septumwas either not formed at all at 5 days of incubation, or was barely noticeableexternally, but then resumed normal morphogenesis. In group 11 embryos (28),in which two fibers were placed in the primitive right ventricle, the distance be-tween them always increased on each subsequent examination (Fig. 3, Table 4).[This also occurred in group III embryos (26), in which two fibers were placed inthe primitive left ventricle, as illustrated in Table 5.

Tritiated thymidine-coating techniques. Six h after fiber placement (4^-day-oldembryos), a major portion of the myocytes of the trabeculae in contact withand adjacent to the fibers were labeled with silver granules (Fig. 4). The myo-

Ventricular septum in chick 21

Table 5. Two fibers in the left ventricle

Embryo no.

VS 110VS 172VS 175VS 176VS202VS213VS234VS247VS270VS271VS341VS345VS346VS349VS352VS358VS362VS363VS364VS371VS372VS375VS378VS643VS645VS651

Age at the time*of operation

4 D -4 D4D +4D +4 D -4D +4 D4 D -4 D -4 D4 D4D +4D +4D +4 D -4 D -4 D4 D4 D +4 D -4 D -4 D -4 D4 D +4 D +4 D -

f

Oh

37530025040020035015010015020020025030012515020025025015010010030040020010075

* 4 D + /— indicates 4

Distance variations betweenthe two i

24 h

750350400650300450300250200350250400450175225300300350225200150400500250150125

ibers (/tm)

48 h

1050575500800400650400350400350450500650350400450400600450425400550800450375325

days plus or minus 3 h.

•\72 h

650775———650575650—700750———650700750—650600775

1050700550—

cardium of the free wall of the ventricle was more heavily labeled than thebodies of the trabeculae, as was the epicardium and pericardium around thesite of the fiber implantation. This was perhaps due to tritiated thymidinescraping off and diffusing into the neighboring area as the fiber entered.

When the fibers were placed anywhere within 200-250 /.im to each side of thefuture ventricular septum, the distance between the labeled cells associated withthe two fibers became shorter 24 h after fiber placement (5-day-old embryos,Fig. 5). The trabeculae within 100-125 ftm of the septum had made contactwith each other. Those to either side of this 100-125 /an wide zone had madeloose contact with septum at their ends projecting into the ventricular cavity,while their bases were still separate from each other. The bases of the trabeculaewere more heavily labeled at this stage than the projecting portions of thetrabeculae. 48 h after fiber placement (6-day-old embryo), the labeled trabeculaewere aggregated densely near the region that would become the crest of the

22 J. Y. HARH AND M. H. PAUL

Fig. 4. Experimental embryonic chick heart at 4£ days (cross-section, from above).Labeled cells (arrows) at and near the site of fiber implantation are seen in trabe-culae, myocardium, and epicardium. Nuclear fast red and picric acid stain, x 70,insets x 200.

smooth septum and upper portion of the trabeculated septum (Fig. 6). Thelabeled cells of the trabeculae originally within 100-125 jtim to each side of thefuture septum were amassed in the central portion of the septum. The labeledcells, on the other hand, in the zone from 125 to 250 jam to each side of thefuture septum were superficially located within the septum, and the bases of thetrabeculae were still more heavily labeled. 72-84 h after fiber placement (7- and7^-day-old embryos), the entire length of all these trabeculae was incorporatedinto the septum (Fig. 7). The labeled myocytes were aligned from the basalportion of the smooth septum to the most apical portion of the trabeculatedseptum, although the silver granules were sparse where the muscular septumand myocardium of the free wall met. The labeled portions of the trabeculaeand myocardium of the ventricular free wall in 4j~day-old embryos wereprogressively shifted into the trabeculae with growth of the ventricles in the5-, 6-, 7-, and 7^-day-old embryos as demonstrated in Fig. 8 and newly formedventricular myocytes were added to below as the ventricles expanded apically.The trabeculae which were immediately outside the zone 400-500 /̂ m wide justdescribed contributed to the ventricular septum at their projecting ends only,and the bases of these trabeculae still remained in the free wall of the ventricle.

Ventricular septum in chick 23

A

Fig. 5. Experimental embryonic chick heart at 5 days (cross-section, from above).Labeled trabeculae are seen in loose contact at the site of the ventricular septum(arrows indicate the labeled cells). A, Section of the apex of the ventricles;B, sectioned through bodies of the ventricles, x 70, x 200.

DISCUSSION

In the present report we have drawn different conclusions about the morpho-genesis of the ventricular septum than others have suggested. Tandler (1912),Waterston (1919), Murray (1919), Takahashi (1923), Kramer (1942), andPatten (1964) suggested that the muscular septum is formed by upward polarizedgrowth of a muscle ridge from the caudal floor of the primary cardiac tube.As Grant (1962) and Goor et ah (1970) have stated, the crest of the septum isstationary and the interventricular foramen is not invaginated by the septum.They demonstrated this by measuring the diameter of the interventricularforamen at each stage of fetal growth. There was no change during the first fewstages, a period of time during which the ventricle had increased several foldin length. Currently this view has been largely disregarded. On the other hand,others have considered more favorably that the septum forms passively by

24 J. Y. HARH AND M. H. PAUL

Fig. 6. Experimental embryonic chick heart at 6 days (from above). Labeledtrabeculae are loosely aggregated in the septum, but still scattered, fb, The sites ofthe fiber implantation. A, Sectioned through the apex of the ventricles; B, cross-section of the bodies of the ventricles; C, cross-section of base of the ventricles.Nuclear fast red and picric acid stain, x 70, x 200.

Ventricular septum in chick 25

w;v/;-*

Fig. 7. Experimental embryonic chick heart at 1\ days (from above). Labeledtrabeculae are completely within the ventricular septum (both smooth and trabe-culated). vsS, Smooth wall of the ventricular septum; vsT, trabeculated ventri-cular septum; Epi, epicardium. A, Cross-section of the apex; B, section throughbodies of the ventricles; C, base of the ventricles. Nuclear fast red and picricacid stain, x 70, x 200.

26 J. Y. HARH AND M. H. PAUL

\

Fig. 8. Diagrammatic drawing of the trabecular aggregation with downwardgrowth of the trabeculae. Note interrelationship of the labeled cells (black symbols).

expansion of the two trabeculated pouches from the primary heart tube on eachside of the interventricular foramen, and that as these pouches become largerand deeper their medial walls coalesce into a common wall. The ventricularseptum, then, is formed by the folding of the two ventricular myocardial walls(Flack, 1909; Frazer, 1932; Streeter, 1948; Grant, 1962; Van Mierop et al1963; Van Mierop & Netter, 1969). Since the septum is formed passively by afolding of the two myocardial walls, the epicardium and the outermost layersof the ventricular myocardium should be aligned most centrally along thelongitudinal axes of the ventricular septum. Therefore, one would not expectto see any communication between the two ventricular chambers at the septum.We have failed to confirm this finding. Instead we found multiple interven-tricular communications from the projecting ends to the basal portions of thetrabeculae during theperiod of septal morphogenesis. These intertrabecular spacesdisappeared only at the final stage of trabecular coaptation. More importantly,we found the morphogenesis of the septum resulted from an active downwardgrowth rather than upward polarized growth. Newly formed ventricularmyocytes were added to the bases of the trabeculae as the ventricles furtherexpanded apically (Fig. 8). De Vries & Saunders (1962) believed that the ventro-medial rotation of the right ventricular limb increases the height of the myo-cardial crest internally, and this becomes the septum as a result of continuedventral infolding and coaptation of the ventricular walls, as well as fusion ofthe trabeculae. This view is essentially a slight modification of Streeter's view,and does not agree with our findings as stated above. Goor et al. (1970) postu-

Ventricular septum in chick 27

lated dual origins of the muscular ventricular septum from the appearance ofits two morphologic structures: a smooth wall above and a trabeculated septumbelow. They then hypothesized that the smooth septum originates by activegrowth of the crest of the smooth wall, which is a process of competitionbetween outward expansion and inward proliferation. It is clearly indicated inthe present study that the septum is of a single developmental origin; thesmooth wall of the muscular septum is formed by the projecting ends of thetrabeculae while the trabeculated septum is formed by the basal portions ofthe trabeculae. This was indicated by the fact that the labeled myocytes areoriginally located throughout the length of the trabeculated septum.

The long dimensions of the developing myocytes in the free wall of theventricular myocardium are parallel to the surface, in a fashion reminiscent ofpavement epithelium. The cells in the trabeculae are oriented along the longaxes of the trabeculae. The increased cell multiplications in the ventricularmyocardial wall result in distribution of part of their cells into the trabeculae.This was shown in this report by the fact that the labeled cells of the commonwall where the septum and myocardium of the free wall met were frequentlyreplaced by cells which were unlabeled (Fig. 8). It is of interest to note that thesedifferences in cell multiplication rate between the septum and myocardial wallcan be seen in early chick embryo studies published earlier (Grohmann, 1961).

The trabecular aggregations in septal morphogenesis are evidenced by theshortening of the distance between the two fibers and between the labeled cellsfrom the site of initial fiber implantation.

The trabeculation of the muscular septum is a form of incomplete coaptationof the trabeculae located near the margins of the 400-500 ^m belt-like zone inthe 4-day-old embryo. The trabeculae located immediately outside this zonebecome the trabeculation of the muscular septum by the fusion of projectingends of the trabeculae into the cavity and their bases remain in the myocardiumof the ventricular free wall.

The authors are indebted to Dr Gerald Odell for his support in this work and to DrsGlenn Rosenquist and Alan Cohen for their laboratory equipment and their interest in ourwork. The Ms Soame Christianson and Lauren Sweeney are acknowledged for their assistancein the preparation of the manuscript. This study was supported by USPHS Grants HD 00091and HL 10191, and Park Ridge (Illinois) United Fund, Inc.

REFERENCES

BELANGER, L. F. & LEBLOND, C. P. (1946). Mineralization of the growing tooth as shown byradio-phosphorus autographs. Endocrinology 39, 309-319.

DE VRIES, P. A. & SAUNDERS, J. B. DE C. M. (1962). Development of the ventricles andspiral outflow tract in the human heart. Carnegie Instn Wash. Contr. Embryo/. 37, 87-114.

FEITELBERG, S. & GROSS, W. (1970). Autoradiography. In Radioactive Nuclides in Medicineand Biology (ed. E. H. Quimby, S. Feitelberg & W. Gross), pp. 338-345. Philadelphia: Leaand Febiger.

FLACK, M. (1909). The heart. In Further Advances in Physiology (ed. L. E. Hill), pp. 34-71.New York: Longmans Green.

28 J. Y. HARH AND M. H. PAUL

FRAZER, J. E. (1932). Development of the heart, and vessels of the anterior part of theembryo. In Manual of Embryology, pp. 306-326. New York: William Wood and Company.

GOOR, D. A., EDWARDS, J. E. & LILLEHEI, C. W. (1970). The development of the inter-ventricular septum of the human heart; correlative morphogenetic study. Chest 58,453-467.

GRANT, R. P. (1962). The embryology of ventricular flow pathways in man. Circulation 25,756-779.

GROHMANN, D. (1961). Mitotische Wachstumintensitat des Embryonalen und FetalenHunchenherzens und ihre Bedeutung fur Entstehung von Herzmissbildungen. Z.Zellforsch. mikrosk. Anat. 55, 104-122.

GROSS, J., BOGOROCH, R., NADLER, N. J. & LEBLOND, C. P. (.195.1). The theory and methodsof the radioautographic localization of radioelements in tissues. Am. J. Roentg. 65,420-458.

HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the development ofthe chick embryo. /. Morph. 88, 49-92.

HARH, J. Y., PAUL, M. H., GALLEN, W. J., FRIEDBERG, D. Z. & KAPLAN, S. (1973). Experi-mental production of hypoplastic left heart syndrome in the chick embryo. Am. J. Cardiol.31, 51-56.

KRAMER, T. C. (1942). The partitioning of the truncus and conus and formation of themembranous portion of the interventricular septum in the human heart. Am. J. Anat.71, 343-370.

MESSIER, B. & LEBLOND, C. P. (1957). Preparation of the coated radioautographs by dippingsections in fluid emulsion. Proc. Soc. exp. Biol. Med. 96, 7-10.

MURRAY, H. A. (1919). The development of the cardiac loop in the rabbit, with specialreferences to the bulboventricular groove and origin of the interventricular septum.Am. J. Anat. 26, 29-39.

PATTEN, B. M. (1964). The heart. In Foundation of Embryology, pp. 545-569. New York:McGraw Hill.

ROSENQUIST, G. C. & DEHAAN, R. L. (1966). Migration of pre-cardiac cells in the chickembryo. Carnegie Instn Wash. Contr. Embryol. 38, 111-121.

SISSMAN, N. J. (1966). Cell multiplication rates during development of the primitive cardiactube in the chick embryo. Nature, Lond. 210, 504-507.

STREETER, G. L. (1948). Developmental horizons in human embryos. Description of agegroups XV, XVI, XVII, and XVIII, being the third issue of a survey of the CarnegieCollection. Carnegie Instn Wash. Contr. Embryol. 32, 133-209.

TAKAHASHI, S. (1923). Notes on the formation of the cardiac septa in the chick. /. Anat.57, 168-177.

TANDLER, J. (1912). The development of the heart. In Manual of Human Embryology, vol. 2(ed. F. Keibel & F. P. Mall), pp. 534-570. Philadelphia: J. B. Lippincott.

VAN MIEROP, L. H. S., ALLEY, R. D., KAUSEL, H. W. & STRANAHAN, A. (1963). Pathogenesisof transposition complexes. I. Embryology of the ventricles and great arteries. Am. J.Cardiol 12, 216-225.

VAN MIEROP, L. H. S. & NETTER, F. H. (1969). Embryology, Ciba Collection 5, 112-130.WATERSTON, D. (1919). The development of the heart in man. Trans. Roy. Soc. Edinb. 52,

257-304.{Received 4 March 1974)