In order to clarify the relationships between the first cleavage plane and the embryonic axes, early cleavage pattern of the fertilized eggs of the starfish, Asterina pectinifera was reexamined. It was ascertained that the polar bodies were formed at the site to which the germinal vesicle had closely located before the initiation of the meiotic division, and that the first cleavage plane passed near this site of polar body formation. While some of the early embryos of this starfish were observed to show various cleavage patterns during early cleavage stage, more than 70 % of the embryos developed according to, so to say, the ‘typical’ cleavage pattern.

Next, horseradish peroxidase (HRP) was injected into one of the blastomeres of the 2-cell- or 8-cell-stage embryos. The embryos were allowed to develop up to either the early gastrula or the early bipinnaria stage and stained to detect the descendants of the blastomere injected with HRP. In early gastrulae still retaining radial symmetry, the activity of HRP injected at the 2-cell stage was found only in one side of the embryo partitioned by one of the symmetrical planes. When one of the four blastomeres lying nearer to the polar bodies at the 8-cell stage was marked with HRP, its descendants constituted one quarter of the anterior part of the gastrula, and descendants of a blastomere opposite the polar bodies were found in the posterior region of the embryo. It was concluded that the animal-vegetal (AV) axis was preexisting in the fertilized egg and that the first cleavage plane contained this primary axis.

In early bipinnariae with their dorsoventral (DV) axes already established, the region of activity of the HRP injected at the 2-cell stage was still demarcated by a plane which passed through the AV axis, but the plane of the boundary had no fixed relation to the DV axis. The results indicate that the first cleavage plane does not necessarily correspond to the median plane of the starfish larva, unlike the case in sea-urchin eggs (Horstadius & Wolsky, 1936). In other words, the DV axis of the starfish embryo is not predetermined in the fertilized egg, and might be established in the course of development through cell-to-cell interactions, while the AV axis is established mainly according to the pre-existing egg polarity.

Establishment of the embryonic axes in early development of echinoderm eggs has been studied by several workers (reviewed by Horstadius, 1973). Using a surgical technique and the famous pigment band as a natural marker, Horstadius (1927) provided evidence that the animal-vegetal (AV) axis pre-exists before fertilization in Paracentrotus eggs. Recently Schroeder (1980) extended this preexistence of the AV axis to other species of sea urchins by employing the jelly canal as another marker of the egg.

Few experimental analyses have been made, however, of the establishment of the secondary axis, that is, the dorsoventral (DV) axis. Difficulty in analysing the properties of the DV axis is due partly to lack of a proper natural marker, but mostly to the highly ‘regulative’ properties of echinoderm eggs: those experiments that involve the separation or removal of some part of the embryo, which has been a successful means for examining ‘mosaic’ type eggs, do not provide definite answers. The crucial experiment would be to observe intact whole embryos, preferably with some sort of artificial markers.

In recent years, it has been demonstrated that the use of horseradish peroxidase (HRP) is an excellent means for tracing cell lineages (Weisblat, Sawyer & Stent, 1978; Hirose & Jacobson, 1979; Batakier & Pedersen, 1982). By employing HRP as a marker, the author could follow the descendants of the HRP-injected blastomeres in starfish embryos. To my knowledge, no systematic experimental study has been made with starfishes to determine either the AV or DV axis of the egg. Since starfish eggs do not form micromeres, the embryonic axis is not expressed until the late blastula stage. This disadvantage of starfish egg as material seems to have been overcome by the use of HRP, as reported in the present paper.

Animals

Asterina pectinifera were collected along the shore of Tokyo Bay in April and Wakasa Bay in September, and kept in aquaria supplied with circulating cold sea water (15 °C).

Oocyte maturation

Excised ovaries were rinsed three times with artificial sea water (A.S.W., Jamarin U, Jamarin Lab., Osaka), and immersed in A.S.W. containing 2×10−6M-l-methyladenine (Kanatani, Shirai, Nakanishi & Kurokawa, 1969) for 20min at 20°C to induce reinitiation of meiosis. After this treatment, oocytes that had undergone germinal vesicle breakdown were shed from the ovaries. These were rinsed three times with A.S.W.

Insemination

Oocytes were inseminated before first polar body formation with a suspension of dry sperm. By inseminating the oocytes during this period, the time of the first cleavage was scheduled according to the time of 1-methyladenine treatment (Kominami & Satoh, 1980), so the time of the initiation of treatment was employed to describe the developmental time. Embryos were cultured in A.S.W. at 20 °C.

Injection of HRP into blastomeres

Injection pipettes were prepared by pulling glass tubes (Ø= 1 mm) over a micro gas burner. The tips were broken with watchmaker’s forceps to make the final tip diameter 3– 5 μm. HRP (type VI; Sigma Chemical Co., St. Louis, MO.) was dissolved at a concentration of 10 % in Ca++-free artificial sea water. This solution was introduced into an injection pipette, and sealed with liquid paraffin. The micropipette was held by a micromanipulator (Model MK-2, Narishige Sci. Inst. Lab., Tokyo) and the solution was injected into the blastomeres, using an injection syringe (Injecting device, Ernst Leitz Ltd., Midland, Ontario) under a microscope. Embryos at the 2-cell or 8-cell stage with an intact fertilization membrane could be penetrated by the injection pipette. An intact fertilization membrane was indispensable for preserving the organization of the cleaved blastomeres (Dan-Sohkawa, 1976). The quantities of solution injected were not precisely determined, but were about 4– 10pl for the 2-cell- and 1– 2pl for the 8-cell-stage embryos.

Enzyme reaction

Early gastrulae (24– 28 h after the induction of maturation, 20 °C) and early bipinnariae (48– 52h) were fixed with 1·5% glutaraldehyde in A.S.W. for Ih then rinsed in 0·1 M-phosphate buffer (pH 6·4) for 30 min at 20 °C. These samples were preincubated for 20 min in 0·3 % 3,3’-diaminobenzidine dissolved in 0·1 M-phosphate buffer then an adequate quantity of 1 % peroxide was added (the final concentration was about 0·01 %). The reaction was continued for 5– 10 min and was stopped by washing the samples twice with 0·1 M-phosphate buffer.

Polar body formation

In immature oocytes the germinal vesicle is located eccentrically, closely apposing the oocyte membrane. The polar bodies are subsequently formed at this point after the reinitiation Of meiosis induced by 1-methyladenine (Shirai & Kanatani, 1980, Fig. 1, A-C). This spatial relationship was not modified by fertilization, as judged from the time-lapse videofilms taken from the initiation of germinal vesicle breakdown through fertilization to the successive formation of the two polar bodies (Fig. 1, D-F).

Fig. 1.

Polar body formation. (A) and (D) Germinal vesicles located eccentrically, closely apposing the oocyte membrane. Arrows indicate the presumed sites of polar body formation. (B) After germinal vesicle breakdown. (C) At the nearest point of the germinal vesicle to the oocyte surface two polar bodies have been extruded (arrowheads). Oocytes were not inseminated. (E) The presumed sites of polar body formation was not modified by fertilization. (F) The first and second cleavage furrows passed near the site of polar body formation. Bar: 100μm.

Fig. 1.

Polar body formation. (A) and (D) Germinal vesicles located eccentrically, closely apposing the oocyte membrane. Arrows indicate the presumed sites of polar body formation. (B) After germinal vesicle breakdown. (C) At the nearest point of the germinal vesicle to the oocyte surface two polar bodies have been extruded (arrowheads). Oocytes were not inseminated. (E) The presumed sites of polar body formation was not modified by fertilization. (F) The first and second cleavage furrows passed near the site of polar body formation. Bar: 100μm.

First cleavage

First cleavage took place about 170 min after the initiation of 1-methyladenine treatment at 20 °C. The plane of the first cleavage is commonly said to pass through the site of polar body formation. To make sure that such a spatial relationship exists, the distance of the site of polar body formation from the first cleavage plane was measured on photographs after the two blastomeres were completely separated by the cleavage furrow (Fig. 2). This distance was less than 20 pm in most cases, and even in extreme cases did not exceed 50μm. Since the shorter diameter (perpendicular to the first cleavage plane) of blastomeres is about 100μm, it can be said that the first cleavage furrow passes near the polar body formation site.

Fig. 2.

Distance of the polar-body-formation site from the first cleavage plane. Ordinate; Number of embryos. Abscissa; Distance of polar bodies from the first cleavage plane (μ m). In more than 90 % of the embryos, this distance was less than 20 gm. Blastomeres of the 2-cell-stage embryos were about 100μ m in shorter diameter.

Fig. 2.

Distance of the polar-body-formation site from the first cleavage plane. Ordinate; Number of embryos. Abscissa; Distance of polar bodies from the first cleavage plane (μ m). In more than 90 % of the embryos, this distance was less than 20 gm. Blastomeres of the 2-cell-stage embryos were about 100μ m in shorter diameter.

Early cleavages

Second cleavage took place perpendicular to the first cleavage plane, and also passed near the polar body formation site. After two divisions, the polar bodies were observed on either one of the four blastomeres. The third cleavage plane was perpendicular to both the first and the second cleavage planes. Frequently the third cleavage was slightly unequal. In this case the blastomeres nearer to the site of polar body formation were smaller than their sister blastomeres. After the fourth cleavage, two layers of blastomeres, each containing eight blastomeres in a circle, were formed (Fig. 3, IA-D). In the fifth cleavage each layer divided into two layers, resulting in four layers of eight blastomeres each. Thereafter the cleavage direction became non-uniform, so it is difficult to describe the organization of the blastulae. This cleavage pattern (Fig. 31) occurred in about 70 % of the embryos, and will hereafter be called the ‘typical’ pattern. In the rest of the embryos, the ‘typical’ pattern was altered as early as the second cleavage. In some embryos the two cleavage planes were oblique with each other (Fig. 311). In extreme cases, the two cleavage axes were perpendicular (Fig. 3III), although the 8-cell-stage embryos apparently had the same arrangement of blastomeres as the ‘typical’ embryos.

Fig. 3.

Various cleavage patterns during the early cleavage stages (from 2-cell to 16-cell stage). IA-ID; The ‘typical’ cleavage pattern occurring in about 70 % of the embryos. The 16-cell-stage embryos consist of two layers of eight blastomeres each. IIA-IID ; The second cleavage plane of one blastomere is oblique to that of the other blastomere, so the arrangement at the 8-cell or 16-cell stage is disturbed. About 20 % of the embryos had this cleavage pattern. IIIA-IIID; The second cleavage plane of one blastomere is almost perpendicular to that of the other blastomere; this case represents about 10 % of the embryos. An embryo of this type appears ‘typical’ at the 8-cell, but at the 16-cell stage the arrangement of blastomeres is again disturbed. Arrows indicate the sites of polar body formation. Bar: 100 μm.

Fig. 3.

Various cleavage patterns during the early cleavage stages (from 2-cell to 16-cell stage). IA-ID; The ‘typical’ cleavage pattern occurring in about 70 % of the embryos. The 16-cell-stage embryos consist of two layers of eight blastomeres each. IIA-IID ; The second cleavage plane of one blastomere is oblique to that of the other blastomere, so the arrangement at the 8-cell or 16-cell stage is disturbed. About 20 % of the embryos had this cleavage pattern. IIIA-IIID; The second cleavage plane of one blastomere is almost perpendicular to that of the other blastomere; this case represents about 10 % of the embryos. An embryo of this type appears ‘typical’ at the 8-cell, but at the 16-cell stage the arrangement of blastomeres is again disturbed. Arrows indicate the sites of polar body formation. Bar: 100 μm.

Distribution of HRP-labelled cells in early gastrulae

a) Injection at the 2-cell stage

Early gastrulae with an archenteron at the vegetal pole were cylindrical in shape, and radially symmetrical around the AV axis. HRP was injected into either blastomere of the 2-cell-stage embryos. Samples were fixed 24– 28 h after the reinitiation of meiosis and observed for the distribution of HRP-labelled cells. Some results are shown in Fig. 4. After the enzyme reaction, labelled cells were stained reddish brown, and could be easily distinguished from the nonlabelled cells. At a glance it is clear that labelled cells and non-labelled cells did not extensively intermingle, and that the boundary between the two groups of cells was rather clear. The distribution of labelled cells in the gastrula shown in Fig. 4A is just the right half of the embryo. Clearly, labelled and non-labelled regions are separated by a plane including the AV axis. Figure 4B shows a gastrula in which the labelled ectodermal region was somewhat smaller than the non-labelled region, whereas the labelled endodermal region was somewhat larger. This indicates a slight shift in the plane of partition from the AV axis. In Fig. 4C, the labelled ectodermal region was somewhat larger than the nonlabelled ectoderm, and the labelled archenteron was somewhat smaller compared with the non-labelled part, the reverse of Fig. 4B. In total, about half the body of the embryo seemed to be labelled. This situation is more clearly seen in Fig. 4D, which is a top view of the same specimen as in Fig. 4C. In most embryos, labelled cells were found only on one side of the embryo with respect to the symmetrical planes of the early gastrula. As the boundary can be thought to coincide with the first cleavage plane, it is concluded that one of the blastomeres of the 2-cell-stage embryo gave rise to half of both the ectoderm and endoderm of the early gastrula. This may also indicate that the first cleavage plane contained the AV axis of the future embryo.

Fig. 4.

Early gastrulae (24– 28 h) following injection of HRP into one of the blastomeres at the 2-cell stage. (A) HRP activity is found only in the right half of this early gastrula. Arrows indicate the boundary between labelled and non-labelled cells. The boundary contains the AV axis of the embryos. (B) Another example in which the boundary has shifted a little to the right of the anterior head region and to the left of the archenteron tip. (C) HRP activity is detected in the left half. (D) Top view of the same embryo shown in C. Bar: 100 μm.

Fig. 4.

Early gastrulae (24– 28 h) following injection of HRP into one of the blastomeres at the 2-cell stage. (A) HRP activity is found only in the right half of this early gastrula. Arrows indicate the boundary between labelled and non-labelled cells. The boundary contains the AV axis of the embryos. (B) Another example in which the boundary has shifted a little to the right of the anterior head region and to the left of the archenteron tip. (C) HRP activity is detected in the left half. (D) Top view of the same embryo shown in C. Bar: 100 μm.

b) Injection at the 8-cell stage

Next, HRP was injected into one of the blastomeres at the 8-cell stage. Unlike those described above (Fig. 31), the cleavage patterns were not so uniform among embryos. Only the embryos that had shown the ‘typical’ pattern at the 4-cell stage were used in this experiment. The half containing the polar-bodyformation site with respect to the third cleavage plane was named the ‘upper’ half and the other half named the ‘lower’ half. HRP was injected into one of the blastomeres of the ‘upper’ half at the 8-cell stage, and the distribution of labelled cells in the early gastrula was observed. Figures 5A and 5B illustrate one of the results. Labelled cells were localized in one quadrant of the embryos distributed from the animal pole down to two thirds of the ectoderm in the AV direction, as could be seen by viewing the embryo from the animal pole. When HRP was injected into one of blastomeres of the ‘lower’ half at the 8-cell stage, labelled cells were found both in the ectodermal region near the vegetal pole and in some part of the archenteron (Figs 5C and 5D). These results indicate that blastomeres in the ‘upper’ half of the 8-cell-stage embryo constituted only the ectoderm of the animal region, and that blastomeres of the ‘lower’ half differentiated into both ectodermal cells near the vegetal pole and endodermal cells that gave rise to the archenteron of the early gastrula. Therefore, the site of polar body formation corresponds to the animal pole (head region, or the most anterior part) of the early gastrula, and its antipode corresponds to the region where gastrulation begins.

Fig. 5.

Early gastrulae (24– 28 h) following injection of HRP into one of the blastomeres at the 8-cell stage. (A) HRP was injected into one of the ‘upper’ blastomeres. Labelled cells are distributed from the anterior (animal) to the posterior pole in two thirds of the ectoderm. (B) The same embryo shown in A. The anterior ectoderm of about one fourth the circumference is labelled. (C) HRP was injected into one of the ‘lower’ blastomeres. HRP activity is found in the posterior ectoderm and in the archenteron. (D) Another example of the same type of injection with the embryo squashed slightly to more easily reveal the distribution of labelled cells. Arrows indicate the boundaries. Bar: 100 μ m.

Fig. 5.

Early gastrulae (24– 28 h) following injection of HRP into one of the blastomeres at the 8-cell stage. (A) HRP was injected into one of the ‘upper’ blastomeres. Labelled cells are distributed from the anterior (animal) to the posterior pole in two thirds of the ectoderm. (B) The same embryo shown in A. The anterior ectoderm of about one fourth the circumference is labelled. (C) HRP was injected into one of the ‘lower’ blastomeres. HRP activity is found in the posterior ectoderm and in the archenteron. (D) Another example of the same type of injection with the embryo squashed slightly to more easily reveal the distribution of labelled cells. Arrows indicate the boundaries. Bar: 100 μ m.

Distribution of HRP-labelled cells in early bipinnariae

a) Injection at the 2-cell stage

HRP-injected embryos developed to the early bipinnaria stage (48– 52 h after the initiation of development) were fixed and stained for HRP. At this stage of development, the DV axis of the embryo could be easily discerned by the presence of the oral opening on the ventral side. The archenteron had differentiated into oesophagus, stomach, intestine and mesodermal tissues (coeloms and mesenchyme cells). Some samples are shown in Fig. 6 and in Fig. 7 with diagrammatic representations. In the example shown in Fig. 6B and in Fig. 7B, labelled cells were found only in the right half of the embryo, and the boundary between the two groups of cells (labelled and non-labelled cells) coincided with the median plane of the early bipinnaria. Each blastomere of the 2-cell-stage embryo thus constituted just the right or the left half, respectively, of the early bipinnaria. However, this situation was observed in only a few cases. In the example shown in Fig. 6B and in Fig. 7B, labelled cells were found only in the ventral half of the embryo. The DV axis of the embryo, in this case, was determined in a direction perpendicular to the first cleavage plane. Directly opposite to the example in Fig. 6B and Fig. 7B, the dorsal half of the embryo (both ectoderm and endoderm) was labelled in the example in Fig. 6C and Fig. 7C. Separation of labelled and non-labelled cells was in many other cases neither left-right nor dorsal-ventral. As shown in Figs 6D, E, F and Figs 7D, E, F, labelled cells were found both in part of the ventral half and in part of the dorsal half. In these cases, the plane dividing the labelled parts from the non-labelled parts of the embryo did not correspond to the median plane. Instead the planes took various angles to the median planes. As the plane containing the boundary between the two groups of cells (labelled and non-labelled) was thought to correspond with the first cleavage plane, it was concluded that the first cleavage plane did not necessarily coincide with the median plane of the early bipinnariae. In other words, the DV axis was determined independently of the first cleavage plane in the course of development.

Fig. 6.

Early bipinnariae (48– 52 h) following injection of HRP into one of blastomeres at the 2-cell stage. (A) Upper half (right half as to the median plane of the embryo) is labelled. The first cleavage plane in this case corresponds to the median plane of the early bipinnaria. (B) Almost all of the ventral half is labelled and the dorsal half is not labelled. The first cleavage plane in this embryo separates the ventral half from the dorsal half. (C) An embryo which is opposite the embryo shown in B. The dorsal half is labelled. (D), (E), (F) In these embryos, the first cleavage planes are oblique to the median plane of the early bipinnariae. Portions of both the ventral and dorsal halves are labelled, but the labelled area is still about half of each embryo. Arrows indicate the boundaries. Notice the labelled cells, especially at the anus. Bar: 100μ m.

Fig. 6.

Early bipinnariae (48– 52 h) following injection of HRP into one of blastomeres at the 2-cell stage. (A) Upper half (right half as to the median plane of the embryo) is labelled. The first cleavage plane in this case corresponds to the median plane of the early bipinnaria. (B) Almost all of the ventral half is labelled and the dorsal half is not labelled. The first cleavage plane in this embryo separates the ventral half from the dorsal half. (C) An embryo which is opposite the embryo shown in B. The dorsal half is labelled. (D), (E), (F) In these embryos, the first cleavage planes are oblique to the median plane of the early bipinnariae. Portions of both the ventral and dorsal halves are labelled, but the labelled area is still about half of each embryo. Arrows indicate the boundaries. Notice the labelled cells, especially at the anus. Bar: 100μ m.

Fig. 7.

Diagrammatic representations of the embryos shown in Fig. 6. Embryos are viewed from the posterior pole with the ventral side upward. Fig. 7A corresponds to Fig. 6A. Fig. 7B corresponds to Fig. 6B, and so on. Dots indicate the distribution of the marker enzyme. D; Dorsal. V; Ventral.

Fig. 7.

Diagrammatic representations of the embryos shown in Fig. 6. Embryos are viewed from the posterior pole with the ventral side upward. Fig. 7A corresponds to Fig. 6A. Fig. 7B corresponds to Fig. 6B, and so on. Dots indicate the distribution of the marker enzyme. D; Dorsal. V; Ventral.

b) Injection at the 8-cell stage

HRP was injected into one of the blastomeres at the 8-cell stage, and embryos were allowed to develop to the early bipinnaria stage. It was expected that only one fourth of the anterior ectoderm would be labelled if HRP were injected into one of the blastomeres of the ‘upper’ half (animal hemisphere), taking into account the results described above. One of these embryos is shown in Figs 8A and 8B. Only the ectoderm of the anterior right half of the ventral side was labelled. It was also expected that about one fourth of the posterior region of the ectoderm and the digestive tract would be labelled if HRP were injected into one of the blastomeres of the ‘lower’ half (vegetal hemisphere). The result represented by Figs 8C and 8D fulfilled this expectation. In addition, the cells of the stomodaeum were labelled in the example shown in Figs 8A and 8B, whereas they were not labelled in the example shown in Figs 8C and 8D. This indicates that the cells of the stomodaeum of the early bipinnaria originate from a blastomere of the ‘upper’ half of the 8-cell-stage embryo, i.e., they originate from the ectoderm of the early gastrula.

Fig. 8.

Early bipinnariae following injection of HRP into one of the blastomeres at the 8-cell stage. (A) HRP was injected into an ‘upper’ blastomere. Ventral view. The anterior ectoderm is labelled. A double arrow indicates the labelled stomodaeum. (B) Same embryo shown in A. Side view. Arrows indicate the boundary between labelled and non-labelled cells. About two thirds of the ectoderm from anterior to posterior and about one fourth the circumference are labelled. (C) HRP was injected into one of the ‘lower’ blastomeres at the 8-cell stage. Ventral view. The posterior ectoderm and the archenteron are labelled. A double arrow indicates the nonlabelled stomodaeum. (D) Side view of the same embryo shown in C. Bar: 100μm.

Fig. 8.

Early bipinnariae following injection of HRP into one of the blastomeres at the 8-cell stage. (A) HRP was injected into an ‘upper’ blastomere. Ventral view. The anterior ectoderm is labelled. A double arrow indicates the labelled stomodaeum. (B) Same embryo shown in A. Side view. Arrows indicate the boundary between labelled and non-labelled cells. About two thirds of the ectoderm from anterior to posterior and about one fourth the circumference are labelled. (C) HRP was injected into one of the ‘lower’ blastomeres at the 8-cell stage. Ventral view. The posterior ectoderm and the archenteron are labelled. A double arrow indicates the nonlabelled stomodaeum. (D) Side view of the same embryo shown in C. Bar: 100μm.

Several papers have reported the usefulness of HRP to trace the fate of blastomeres (Weisblat et al. 1978; Hirose & Jacobson, 1979; Batakier & Pedersen, 1982). HRP turned out to be useful for marking blastomeres in starfish embryos also. As shown in Fig. 4, for example, HRP injected into one of the blastomeres at the 2-cell stage was localized on either side of a symmetrical plane of the early gastrula. This indicates that HRP molecules have not moved from the descendent cells of the HRP-injected blastomere to descendants of another blastomere (Tupper & Saunders, Jr, 1972). Blastomeres of the embryo did not intermingle much during early embryogenesis at least up to the early gastrula and in fact these conditions were not modified up to the early bipinnaria stage except for the migration of mesenchyme cells.

The data presented here clearly show that the first cleavage plane contains the AV axis, which becomes the anteroposterior axis in the early gastrula. It was also ascertained that the ‘upper’ blastomeres in reference to the site of the polar bodies formed the anterior part of the gastrula ectoderm and that the ‘lower’ blastomeres formed the posterior ectodermal region and the archenteron (Figs 5 and 8).

HRP-injected, ‘upper’ blastomeres of the 8-cell-stage embryos formed approximately the anterior two thirds of the ectoderm in the early gastrula (24– 28 h), or in early bipinnaria (48– 52 h). The ‘lower’ blastomeres of the 8-cell-stage embryos formed the posterior third of the ectoderm and the archenteron. The 32-cell-stage embryo consists of four layers of eight blastomeres each. Of these four layers, the ‘lower’ two are derived from the ‘lower’ blastomeres of the 16-cell-stage embryo. The present results indicate that only the lowest layer of blastomeres at the 32-cell stage may form the digestive tract.

The most significant result obtained in these experiments is that the first cleavage plane does not necessarily contain the DV axis of the starfish larva (Figs 6 and 7). In amphibian eggs, the first cleavage plane, which becomes the future median plane of the embryo, passes through the sperm entrance point and the grey crescent that is formed opposite the sperm entrance point. Therefore, the entrance of a spermatozoon into an egg is the critical event in establishment of the DV axis before the first cleavage (Ancel & Vintemberger, 1948). Horstadius & Wolsky (1936) concluded that the DV axis pre-exists at the 1-cell stage of the fertilized eggs of the sea urchin, on the basis of their experiments involving isolation and recombination of vitally stained blastomeres. The observation of Horstadius (1925) on Asterias gibbosa is the sole prior information on the origin of the DV axis in starfish eggs. Some of this species are reportedly oval in shape. Taking advantage of this irregularity, he found that the first cleavage furrow separates dorsal from ventral side in the larvae of that starfish. However, no similar study has ever been reported using spherical eggs. HRP-injection methods as adopted in the present work seem to have offered a clear answer. The DV axis of the starfish larva is probably established during early embryogenesis by means of cell-to-cell interactions.

Determination of the AV axis, on the other hand, may be traced to the stage of polar body formation, since the present study revealed the fixed relationship between the site of polar body formation and the eccentric position of the germinal vesicle, a relationship which is not affected by fertilization. Possibly, the AV axis could be determined even in immature oocytes. A similar conclusion was reached using vital staining methods (Shirai & Kanatani, 1980). Each blastomere isolated from the 16-cell-stage embryo is known to be totipotent, that is, to be able to construct a dwarf but morphologically normal bipinnaria (Dan-Sohkawa & Satoh, 1978). The problem of the embryonic axis in relation to such regulative development would be whether each blastomere develops by newly establishing an AV axis or develops according to its own polarity possibly laid down along the AV axis of the whole embryo before isolation.

Though it turns out to be clear that there is no fixed correlation between the first cleavage plane and the median plane of the larva, the role of sperm entrance in establishing the DV axis remains to be resolved. The entrance of a spermatozoon may affect the embryonic axes determination just as in amphibian embryos (Ancel & Vintemberger, 1948). This is a debated subject for starfishes.

I wish to thank Associate Prof. Takeshi Yanase, Osaka Kyoiku Univ., for affording us opportunities to utilize glass tube puller and also thank Prof. Mitsuki Yoneda for his technical advice during the course of this study and critical reading of the manuscript.

Ancel
,
P.
&
Vintemberger
,
P.
(
1948
).
Recherches sur le déterminisme de la symétrie bilatérale dans l’oeuf des amphibiens
.
Bull. Biol. Fr. Belg. Suppl
.
31
,
1
119
.
Baeakier
,
H.
&
Pedersen
,
R. A.
(
1982
).
Allocation of cells to inner cell mass and trophec-toderm lineages in preimplantation mouse embryos
.
Devi Biol
.
90
,
352
362
.
Dan-Sohkawa
,
M.
(
1976
).
A ‘normal’ development of denuded eggs of the starfish, Asterina pectinifera
.
Devi, Growth and Differ
.
18
,
439
445
.
Dan-Sohkawa
,
M.
&
Satoh
,
N.
(
1978
).
Studies on dwarf larvae developed from isolated blastomeres of the starfish, Asterina pectinifera
.
J. Embryol. exp. Morph
.
46
,
171
185
.
Hirose
,
G.
&
Jacobson
,
M.
(
1979
).
Clonal organization of the central nervous system of the frog. I. Clones stemming from individual blastomeres of the 16-cell and earlier stage
.
Devi Biol
.
71
,
191
202
.
Hôrstadius
,
S.
(
1925
).
Entwicklungsmechanische Studien an Asterias gibbosa Forbes
.
Ark Zool. B
.
17
, No.
6
,
1
6
.
Hôrstadius
,
S.
(
1927
).
Studien fiber die Determination bei Paracentrotus lividus LK
.
W. Roux Arch. EntwMech. Org
.
112
,
239
246
.
Hôrstadius
,
S.
(
1973
).
Experimental embryology of echinoderms
.
Oxford
:
Clarendon Press
.
Hôrstadius
,
S.
&
Wolsky
,
A.
(
1936
).
Studien über die Determination der Bilateralsymmetrie des jungen Seeigelkeimes
.
W. Roux. Arch. EntwMech. Org
.
135
,
69
113
.
Kanatani
,
H.
,
Shirai
,
H.
,
Nakanishi
,
K.
&
Kurokawa
,
T.
(
1969
).
Isolation and identification of meiosis inducing substance in the starfish Asterias amurensis
.
Nature
221
,
273
274
.
Kominami
,
T.
&
Satoh
,
N.
(
1980
).
Temporal and cell-numerical organization of embryos in the starfish, Asterina pectinifera
.
Zool. Mag
.
89
,
244
251
.
Shirai
,
H.
&
Kanatani
,
H.
(
1980
).
Effect of local application of 1-methyladenine on the site of polar body formation in the starfish oocytes
.
Devi, Growth and Differ
.
22
,
555
560
.
Schroeder
,
T. E.
(
1980
).
Expression of the prefertilization polar axis in sea urchin eggs
.
Devi Biol
.
79
,
428
443
.
Tupper
,
J. T.
&
Saunders
,
JR
, J. W
. (
1972
).
Intercellular permeability in the early Asterias embryos
.
Devi Biol
.
27
,
546
554
.
Weisblat
,
P. A.
,
Sawyer
,
R. T.
&
Stent
,
G. S.
(
1978
).
Cell lineage analysis by intracellular injection of a trace enzyme
.
Science
202
,
1295
1298
.