ABSTRACT
Several experiments done in our laboratory make it likely that in the egg of Limnaea stagnalis there is a cortical morphogenetic field (Raven, 1949, 1952, 1966). One of us (Raven, 1963,1964,1967) has studied the origin of this morphogenetic field. In the newly laid egg cell there is a vegetative pole plasm, occupying a sector of about 110 degrees with its apex near the centre of the egg. It is situated somewhat obliquely with respect to the longitudinal axis of the first maturation spindle. Moreover, a circle of six lenticular subcortical patches of cytoplasm are found in the equatorial region of the egg. These ‘subcortical accumulations’ (SCA) are arranged according to a regular pattern. Four of them are situated close together on one side, occupying about 180 degrees of the egg circumference; two somewhat larger SCA lie on the opposite side. The SCA, together with the obliquity of the vegetative pole plasm, define a pattern which is polar, dorsoventral and nearly symmetric, though a slight asymmetry seems to be present.
The following facts make it very likely that the pattern of cytoplasmic differentiations of the recently laid egg is important for cell determination at later development: (1) the plane of symmetry of this pattern coincides with the median plane of the future embryo (Raven, 1967); (2) the egg substances contained in the vegetative pole plasm and in the SCA are distributed among the cleavage cells according to a definite programme (Raven, 1946, 1967); (3) a disturbance of this distribution by centrifugation during early cleavage leads to abnormal development (Raven, 1964).
This cytoplasmic pattern arises by ooplasmic segregation during the passage of the egg cell through the female genital tract of the parent. Centrifugation experiments have shown that the normal course of ooplasmic segregation is apparently controlled by factors residing in the egg cortex (Raven, 1964, 1967; Raven & van der Wai, 1964). Therefore, it must be assumed that the cytoplasmic differentiations of the uncleaved egg arise under the influence of a pre-existent mosaic pattern of the molecular structure of the cortex. This cortical pattern is probably formed during oogenesis, through interactions between the oocyte and the surrounding structures of the gonad. This was concluded from the fact that the arrangement of the SCA of the oviposited egg agrees with the configuration of the follicle cells surrounding the oocyte in the gonad, while the position of the vegetative pole plasm corresponds to the part of the egg surface formerly applied to the gonad wall (Raven, 1963, 1967). The hypothesis was put forward that the ‘blue-print information’ (Raven, 1958) in the egg cortex is transmitted from the parent to the offspring by way of the follicle, whose structure is, so to speak, ‘imprinted’ upon the egg during oogenesis.
Since most of these relationships have been studied in the single species L. stagnalis, the question arises whether these conclusions have a more general bearing. It seems appropriate, therefore, to test them by the study of the eggs of other species.
The cortical morphogenetic field does not only direct the orderly displacement of the egg substances during ooplasmic segregation, but also controls the positions and directions of nuclei and spindles, and thereby the place and direction of cleavage furrows (Raven, 1964). In this way the two processes of ooplasmic segregation and cell division are bound together into a unified pattern, ensuring the orderly distribution of the egg components among the cleavage cells.
In this connexion the question becomes important whether the asymmetry of cleavage and of further development in gastropods is due to an asymmetry of the cortical field. It has long been known that the asymmetry of coiling of the shell is correlated with the asymmetry of spiral cleavage, the cleavage of sinistral species of snails being the mirror image of that of dextral snails. That the same regularity holds within a species has been shown by one of us (Ubbels, 1966).
In L. peregra two races exist, one showing dextral coiling and the other sinistral coiling of the shell. The cleavage pattern of eggs from which dextrals arise is the mirror image of that shown by sinistrals. Thus, as in other gastropods, asymmetry of coiling of the shell is correlated with the asymmetry of spiral cleavage. According to Sturtevant’s (1923) hypothesis, which is based on the work of Boycott & Diver (1923; cf. Boycott, Diver, Garstang & Turner 1931) the direction of coiling of the shell in L. peregra is determined by one pair of Mendelian factors, the allele for dextral coiling being dominant. Moreover, the direction of coiling does not depend on the animal’s own genotype, but on the genotype of the female parent. Therefore the asymmetry of the future embryo must be laid down in the egg during oogenesis (Sturtevant, 1923).
If the asymmetry of the cortex plays a part in the determination of the cleavage pattern it may be expected that the cortical patterns of eggs derived from genetically dextral and sinistral L. peregra will mirror each other. If the cortical patterns in their turn are related to the topography of the elements surrounding the growing oocyte, the follicle cell patterns of the two races may be expected to be different, and to mirror each other too. Therefore the arrangement of follicle cells in genetically dextral and sinistral L. peregra was studied.
MATERIALS AND METHODS
Ovotestis material derived from genetically dextral (D) and sinistral (L) individuals was fixed in Zenker’s or Bouin’s fluid, embedded in paraffin, sectioned at 4μ and stained with azan, or according to the ‘pikro-blauschwarz-kernecht rot’ procedure. The follicle cell patterns in the two races were studied with the aid of wax reconstructions. The final magnification was about x 1100. The nuclei and the cell bodies of the follicle cells, and the area of attachment of the oocyte cell membrane to the basal membrane of the acinus were indicated in each model.
Fig. 1 shows the relationship between oocyte size (expressed as the weight of the corresponding wax reconstruction) and the number of follicle cells. It appears that during the first phase of slow growth the oocyte is gradually surrounded by an increasing number of follicle cells. They are presumably recruited from the cells of the germinal epithelium. Cell divisions were never observed in follicle cells of Limnaea, while cells which could be considered as transitional in their morphologic appearance between cells of the germinal epithelium and follicle cells were often encountered in the neighbourhood of the oocytes (Ubbels, 1968). When rapid growth begins, the inner layer of follicle cells immediately adjoining the oocyte is completed. It appears that this consists in L. peregra of seven or eight cells as a rule; six cells are found in a number of smaller follicles, whereas occasional follicles with nine elements do occur.
Relationship between size of the oocyte and number of follicle cells. + = oocytes from genotypically sinistral snails; ○ = oocytes from genotypically dextral snails.
The fact that the number of follicle cells varies in this species thwarts the analysis of follicle cell pattern, as different follicle classes have to be treated separately. It is further complicated by the fact that the distribution of ‘dextral’ and ‘sinistra!’ follicles among the various classes is rather uneven. Of the full-grown follicles studied in detail (not considering the 6-cell follicles, which are generally smaller), among those from genetical dextrals there are nine follicles with seven cells, two with eight and two with nine cells; among the sinistrals two with seven cells, nine with eight cells and one with nine cells (cf. Table 1).
Owing to variations in the shape of the oocyte and the follicle cells a direct spatial comparison of the various follicle cell patterns proved to be difficult. Therefore in each model an imaginary plane was brought through as many follicle cell nuclei as possible. The nuclei which were not included in this plane were projected upon it. The distances between the centres of the successive nuclei or of their projections were measured in a clockwise direction as seen from the apical side. All distances were converted into arcs of a circle with a circumference equal to the sum of the distances.
In Table 1 the positions of the centres of the follicle cell nuclei are expressed in degrees of a circle, arbitrarily taking the position of one of the nuclei as zero.
If the follicle cells within a certain class are arranged according to a pattern, it must be possible to rotate the models in such a way that their patterns coincide. Since there is no fixed point of reference, several possible combinations have to be tried out. This has been done by means of a computer by one of us (J. J. B.). The following considerations determined the method of approach of the problem.
Estimation of patterns
It has been mentioned above that the available material can be classified according to the race of the animal (dextral or sinistral) and to the number of follicle cells (varying from 6 to 9), yielding eight classes in all. Let us consider the problem of determining the existence of a pattern common to the members of a single class. If such a pattern exists, the follicle cells in different oocytes of the same class must occupy corresponding positions, although, of course, a certain amount of individual variability is admissible. We will, therefore, treat the actual positions of corresponding follicle cells as random variables, and define the common pattern as the set of their population means. The problem is how to obtain an estimate of this pattern from the data in Table 1. The difficulty lies in the fact that the oocyte possesses no fixed point of reference that would enable us to establish with certainty which follicle cells are in correspondence. This question has to be decided by means of the data themselves, and they are by no means unambiguous in this respect. Therefore, a criterion is needed leading to a decision that is optimal in some sense. Minimum variance seems to be a reasonable choice. In the following we will explain how this criterion is applied.
Suppose a class contains n oocytes with p follicle cells each. If the way in which the cells correspond is somehow fixed, they can be numbered (corresponding cells receiving the same number), and the np observations can be arranged in a table with n rows and p columns. The following notation will be used:


Comparison of D- and L-patterns
The same minimum variance criterion can be used to decide between hypotheses regarding the relations between patterns in different classes. To this end the classes are combined to a single unit and minimum variance arrangements are determined, corresponding to the various hypotheses to be compared. Once again the lowest value of the variance indicates which of the hypotheses is to be preferred.
In particular, it can be investigated in this way whether the difference in D- and L-asymmetry is reflected in the follicle cell patterns. If the comparison is restricted to oocytes with the same number of follicle cells, the hypotheses to be compared are simple alternatives: either D- and L-patterns are identical, or one is the mirror image of the other. To choose between these alternatives, the minimum variance arrangement is determined for D- and L-oocytes combined, first with the order of the follicle cells unchanged and then with this order reversed in one of the two types. The first or the second alternative is accepted according to whether the lowest value of the variance is found in the first or in the second case.
The use of this mathematical criterion to decide between two biological possibilities emphasizes the necessity of checking its reliability in some way. The changes in value of the variance caused either by renumbering or by reversing the order of the follicle cells are generally small compared with this value itself (cf. Tables 2 and 3 in the next paragraph). Consequently, it is impossible to verify the conclusions statistically, since in most cases the differences would not be significant. It is, however, possible to check the reliability of the criterion by applying it to cases where the result it should produce is known beforehand.
To this end the method described in the previous paragraph for a single class, consisting of oocytes of the same type and with the same number of follicle cells, is extended slightly. Each time a new oocyte is added, it is tried not only with the follicle cells in their original order but also with this order reversed and, if necessary, the order is fixed in the latter state. The criterion can be considered as satisfactory if in the final arrangement the majority of oocytes are in agreement with regard to the order of their follicle cells.
RESULTS
The computations sketched in the foregoing paragraphs were carried out on an Electrologica X8-computer provided with an ALGOL-processor. The programs, written in ALGOL, will not be discussed here but they are available on request.
(a) Reliability of the method
The results obtained by checking the method in the way described above are presented in Table 2. For each class, with the exception of the class 9L which contains only one oocyte, two values of the minimum variance are recorded. The first value holds if the order of the follicle cells is fixed, the second value if this order can be either maintained or reversed. In the columns headed by a + or — sign are shown the numbers of oocytes in which the original order of the follicle cells in the final arrangement is maintained or reversed, respectively.
Table 2 shows that the agreement between oocytes in the minimum variance arrangement with variable order of follicle cells is satisfactory. In the larger classes 7 D and 8L the majority of oocytes (7 or 8 out of 9) are in agreement. With the smaller classes, containing only two or three oocytes each, there is agreement in two cases (6D and 7L), disagreement in two cases (6L and 8D), while in one case (9D) the question remains undecided because the difference in variance with respect to the fixed order arrangement is very small. All things considered, these results seem to justify the use of the criterion for purposes of comparison.
(b) Comparison of D- and L-patterns
The results obtained by combining D- and L-oocytes with the same number of follicle cells are presented in Table 3. For each number of follicle cells three values of the minimum variance are recorded. The first value holds if for D- and L-oocytes both the original order of the follicle cells is maintained, the second value if this order is reversed in one of the two types, and the third value if this order can be maintained or reversed in each individual oocyte of both types. The numbers of oocytes concerned are shown in the same way as in the previous table.
It is difficult to draw any definite conclusions from Table 3 because the results do not all point in the same direction. The best way will be to consider the four classes separately.
In the class of oocytes with six follicle cells a lower value of the variance results if the two types are taken in opposite order. The difference, however, is small: 159 against 167. Moreover, there is no complete agreement between the L-oocytes since the variance is further reduced to the value 149 if only two of them are reversed while the third one is taken in the same order as the D-oocytes.
The only appreciable difference is found in the class of oocytes with seven follicle cells. Here the arrangement with both types in the same order is clearly preferred. Reversing the order in the L-oocytes increases the variance from 205 to 270. Leaving both possibilities open for each individual oocyte results in a still smaller value 187, but even then the two L-oocytes are in agreement with the majority of the D-oocytes.
The class of oocytes with eight follicle cells shows the same picture as that with six cells. There is a very small difference (114 against 119) in favour of the arrangement of the two types in opposite order, but here too the variance assumes a lower value (100) if only one of the two D-oocytes is reversed while the other one is taken in the same order as the L-oocytes. In the class with nine follicle cells the number of oocytes as well as the difference between the values of the variance is too small to give an indication in either direction.
Taking into account all facts listed above, it may be concluded that, although the data are by no means unanimous, there is more evidence in favour of the hypothesis that D- and L-patterns are identical. The estimates given below are based on this assumption.
(c) Estimates of patterns and optimal arrangements
In view of the last conclusion, oocytes of both types with the same number of follicle cells were taken together with all cells in the same order. Optimal arrangements and estimates of patterns were computed for each number of follicle cells. They are presented in Tables 4 and 5. To give an idea of the accuracy of the estimates, standard deviations of the means are likewise recorded in Table 5, where they are indicated by the symbol S.
Description of the follicle cell pattern
The average positions of follicle cells, given in Table 5, determine the patterns of 6-, 7-, 8- and 9-celled follicles respectively. The question arises whether there is a relationship between these patterns. If, as suggested by Fig. 1, the number of follicle cells gradually increases by the recruitment of cells from the germinal epithelium, one may expect that the additional follicle cells will be intercalated between the cells already present. The latter will in general keep their positions, perhaps with the exception of the immediately adjacent cells moving up a short distance to make room for the newcomer. To establish the relationship between follicles with different cell numbers, the four patterns given in Table 5 must be rotated in such a way that the average positions of the cells show the best correspondence to these expectations. This results in the positions given in Table 6 and the corresponding Fig. 2.
Average follicle cell positions in follicles of different cell number. The average positions of follicle cells (1–6) in 6-cell follicles, as seen from the animal pole, are indicated in the inner circle. The cells of 7-, 8- and 9-cell follicles, respectively, have been put in the outward following circles in matching positions.
Average follicle cell positions in follicles of different cell number. The average positions of follicle cells (1–6) in 6-cell follicles, as seen from the animal pole, are indicated in the inner circle. The cells of 7-, 8- and 9-cell follicles, respectively, have been put in the outward following circles in matching positions.
In a previous publication (Raven, 1963) the follicle cells of L. stagnalis have been numbered in a counter-clockwise direction. To preserve a certain uniformity of treatment, we have done the same in the present case; this does not imply, to be sure, that cells with the same number in the two species are considered homologous. The cells of the six-celled follicle of L. peregra are numbered from 1 to 6, starting with the cell at the left end of the largest gap in the series and proceeding counter-clockwise (Fig. 2). With the increase in follicle cell number, cell 7 is inserted between 4 and 5, cell 8 between 1 and 2, and cell 9 between 1 and 6, but nearer to the latter. These numbers have also been placed above the columns of Table 6. It is evident both from the figure and the table that identical cells in general have constant positions within about 10 degrees (cf. columns 6, 5, 7 and 4 of Table 6), with the exception of cells 1 and 2, which move apart to make room for cell 8, and cell 3, taking part in this movement to a lesser extent. This is in agreement with the above-mentioned expectations.
While Fig. 2 gives the average positions of the cells, in Figs. 3 and 4 the individual values of Table 4, but readjusted so as to correspond with the average positions of Table 6 and Fig. 2, have been inserted for the follicles of genetical sinistrals and dextrals, respectively. It appears from these figures that, although the positions of corresponding cells in different follicles exhibit a certain amount of scatter, for most cells they are distinctly clustered around the average values. The areas occupied by adjacent cells can in most cases clearly be delimited against each other, and hardly overlap. This argues for the accuracy of the pattern. Moreover, in comparing Figs. 3 and 4, it is evident that L- and D-follicles do not differ from each other. This confirms the conclusion reached above that L- and D-patterns are identical rather than mirrored.
Individual positions of follicle cells in genetical sinistrals. The plan of the figure and the notation of the cells correspond to Fig. 2.
Subcortical accumulations in the eggs ofL. peregra
In a previous paper (Raven, 1963) the conclusion was reached that the arrangement of the subcortical accumulations (SCA) of the oviposited egg of L. stagnalis reflects the positions of the follicle cells surrounding the growing oocyte in the gonad. In this connexion it appeared important to investigate whether SCA can also be distinguished in the uncleaved eggs of L. peregra, and if their arrangement can be related to the pattern of follicle cells in this species too.
To this end 106 eggs of L. peregra, ranging from anaphase of the first maturation division to anaphase of the second maturation division, have been studied. These eggs had been preserved in 1963, to serve for another investigation, and were probably from genetical sinistrals. They had been fixed in Bouin’s fluid, sectioned at 5 μ, and stained with iron haematoxylin and erythrosin.
A study of the sections shows that in these eggs distinct SCA can be recognized. They have a similar appearance and localization as in L. stagnalis, occupying more or less lens-shaped areas immediately beneath the plasma membrane in the equatorial and vegetative regions (Fig. 5 A, B). Their boundary against the internal cytoplasm is rather irregular, and indented by the vacuoles surrounding y-granules. It is often rather difficult to establish their outline, as they gradually thin out beneath the plasma membrane. In such cases the separation between neighbouring SCA may also be rather vague.
A. Egg P 52-2. Extrusion of first polar body. Egg with six SCA, of which three are visible (c, d, e). Magnification x 560. B. Egg P52-4. Telophase of first maturation division. Egg with six SCA, of which five are visible (a-e; e only marginally sectioned). Magnification × 560. C. Egg P52-4. SCA c-e at higher magnification ( × 875). D. Egg 74-7. Anaphase of second maturation division. Egg with seven SCA. SCA e and f at higher magnification; e marginally sectioned. Magnification ×875.
A. Egg P 52-2. Extrusion of first polar body. Egg with six SCA, of which three are visible (c, d, e). Magnification x 560. B. Egg P52-4. Telophase of first maturation division. Egg with six SCA, of which five are visible (a-e; e only marginally sectioned). Magnification × 560. C. Egg P52-4. SCA c-e at higher magnification ( × 875). D. Egg 74-7. Anaphase of second maturation division. Egg with seven SCA. SCA e and f at higher magnification; e marginally sectioned. Magnification ×875.
The SCA consist of a ‘dense’ cytoplasmic matrix, staining brownish in iron haematoxylin-erythrosin preparations. Dense masses of delicate basophil granules are embedded in this matrix; they stain blue-black with iron haematoxylin. The mitochondria, occurring in the animal pole plasm, around the maturation spindle and asters, and in the cytoplasmic meshes between the vacuoles, exhibit the same staining, but they are coarser and probably rodshaped, whereas the SCA-granules seem to be mainly spherical. The latter are therefore probably not mitochondria. These granules are accumulated for the greater part immediately beneath the plasma membrane in masses often several layers thick, but they may partly also extend inwards in the triangular partitions between adjacent vacuoles (Fig. 5C, D). The βt- and γ-granules of the protein yolk are lacking in the SCA, at least in uncleaved eggs. It is evident that the composition of the SCA agrees with that found in L. stagnalis (Raven, 1967).
In order to study the positions of the SCA, graphical reconstructions were made of all eggs. The method employed was the same as that used for the eggs of L. stagnalis (Raven, 1967). As in that investigation, the positions of the SCA established in the original drawings were transposed by projection on to an idealized equatorial plane, on which the parts of the SCA extending into the vegetal hemisphere were drawn as they would be seen by looking at the egg from the direction of the vegetative pole. By the method outlined in the abovementioned paper, these diagrams were brought into corresponding positions by establishing a system of common meridians, and the co-ordinates of the SCA of each egg in a group were measured. From the median values of these measurements a ‘median projection’ of the group was constructed.
It soon appeared that the 106 eggs studied fell into two groups: in 72 eggs six SCA could be distinguished, whereas the remaining 34 eggs had seven SCA.
Originally the eggs were subdivided into age classes (anaphase of first maturation division, etc.), and median projections of all eggs with the same number of SCA within a single class were constructed. It appeared, however, that no significant changes in the positions of the SCA with age occurred during the period considered (anaphase of first to anaphase of second maturation division). Therefore, the eggs of all age classes with the same number of SCA could be taken together, and a common median projection of them constructed.
Fig. 6 shows the median projection of all (72) eggs with six SCA. The SCA are arranged in a circle beneath the equator of the egg. They are not evenly spaced, however, but on one side there is a wider gap between two adjacent SCA, while on the opposite side at least the centres of two SCA are somewhat wider apart than elsewhere. If we draw a line bisecting these two angles, this defines a meridian plane, with respect to which the centres of the individual SCA are approximately symmetrically arranged. The only apparent asymmetry is in the size of the SCA denoted by c and d in the drawing, d being much more extended in a latitudinal direction.
Limnaea peregra. Median projection of 72 eggs with six SCA. Diagram of vegetal hemisphere; vegetative pole in centre, a-f: the six SCA. Arrows 1-6: corresponding positions of cells 1-6 of 6-cell follicles. Interrupted line: approximate plane of symmetry.
The median projection of the (34) eggs with seven SCA is shown in Fig. 7. Again the SCA are arranged in a subequatorial circle, but their distribution is now clearly asymmetric. If we again draw the meridian plane bisecting the widest gap in the series, three SCA are situated on one side of this plane and four on the other side, though slightly overlapping into the other half. Between the two groups there is a somewhat larger gap than between adjacent SCA of the same group, which may even partly overlap in their latitudinal extension.
Limnaea peregra. Median projection of 34 eggs with seven SCA. Arrows 1-7: positions of cells 1-7 of 7-cell follicles. Interrupted line: meridian plane bisecting widest gap between SCA.
DISCUSSION
Within the limitations of the method used and of the restricted material to which it has been applied, the results described above seem to lead to the conclusion that the follicle cells in egg follicles of L. peregra are not arranged haphazardly around the oocyte, but according to a definite pattern. The following arguments for this conclusion can be adduced:
In the execution of the computations, for every class of follicles (classified according to the race of the animal and the number of follicle cells) many cycles with widely differing sequences of the individual follicles were performed in order to find the arrangement with the lowest variance. It appeared that an identical configuration was found for a majority of the cycles.
If the arrangement of the follicle cells is haphazard, one can expect that a substantial reduction of the variance can be attained by permitting the reversal of the order of the cells in part of the follicles. It appears, however, that this expectation is not fulfilled; in general, the original order of the follicle cells was maintained, and the reversal of some follicles gave only a slight reduction of the variance (cf. Tables 2 and 3).
The patterns resulting from the computations in the various classes show a great resemblance. Not only does it seem probable that the patterns of L- and D-follicles with the same number of cells are identical rather than reversed, but the patterns of follicles with different cell numbers can be arranged in such a way that they conform to reasonable assumptions concerning their mutual relationship (Fig. 2).
The most convincing evidence on the reality of the pattern results from the inspection of Figs. 3 and 4, in which the positions of the cells in the individual follicles are summarized. It appears from these figures that: (1) corresponding cells in different follicles of all size classes are situated in circumscribed areas, which hardly overlap; (2) the single values within an area show a certain amount of scatter, but most of them are distinctly clustered in a certain part of the area; (3) the patterns of L- and D-follicles obtained in this way are superimposable; (4) in both patterns there is a corresponding region of 40-50° of the circumference which is entirely free of follicle cells.
If one takes into account that the recognition of the species-specific pattern of the follicle is bound to be obscured, not only by the normal variability inherent in all biological structures, but also by a summation of distortions caused and errors committed in all preceding phases of the investigation (fixation and sectioning, drawing and making of reconstructions, measurement of cell distances), it is surprising that the main characteristics of the pattern are clearly expressed in these figures. The possibility that such a result could have been produced by the mere process of ordering of follicles with random arrangement of the cells, by the criterion of minimum variance as employed in this investigation, seems rather far-fetched. In this connexion it is important to note that the afore-mentioned errors could not have been biased by anticipation of the expected result, since the latter appeared only afterwards as outcome of the computations. Therefore, it is believed that the resulting pattern conforms to reality.
The pattern of follicle cells in L. peregra is not identical with that found in stagnalis (Raven, 1963), even as regards the 6-cell follicles of the former species. However, a certain general resemblance between the two patterns cannot be denied. In both, the arrangement of the cells is not radially symetrical. On one side of the follicle, the cells are more narrowly spaced than on the other; in L. peregra there is even a wide gap between the cells on one side. Therefore, one can distinguish, in addition to the polar axis connecting the centre of the basal surface with the free pole of the oocyte, a second axis at right angles to the first. In L. stagnalis, it has been proved that this ‘dorsoven-trality’ of the egg follicle coincides with that of the later embryo (Raven, 1967). It remains to be determined whether this also holds for L. peregra.
In L. stagnalis, the pattern of follicle cells is nearly symmetrical with respect to the plane containing the two axes, though a slight asymmetry seems to exist. In L. peregra, in 6- and 8-cell follicles the arrangement of the cells is also roughly symmetrical; in 7- and 9-cell follicles, on the other hand, there is a distinct asymmetry. This asymmetry is the same in L- and D-follicles. Our expectation that the follicle cell patterns of the two races of L. peregra mirror each other has therefore not been fulfilled. Apparently, the determination of the asymmetry of the future embryo (of which we know that it probably occurs during oogenesis) does not take place by way of the asymmetry of the follicle cell pattern, but according to some other mechanism.
In a recent book Waddington (1966) states, referring to our work, ‘. .it has been found that the pattern of asymmetry.. .is reversed for left-handed and right-handed forms’. Though we have alluded to the present work in some previous publications, we have never made any definite statement to this effect. It now appears that this view can no longer be upheld.
In L. stagnalis, in uncleaved eggs six subcortical accumulations (SCA) have been found, the arrangement of which duplicates the pattern of follicle cells previously surrounding the oocyte in the gonad (Raven, 1963, 1967). Since the SCA arise only after ovulation by the accumulation of certain components of the ooplasm beneath particular regions of the plasma membrane, and because their pattern is restored after the displacement of these components by centrifugal force, it was concluded that it reflects a pre-existent mosaic pattern of the egg cortex, which in its turn is formed during oogenesis in correspondence with the structures surrounding the oocyte.
It has now been found that similar relationships occur in L. peregra. Uncleaved eggs of this species have either six or seven SCA. They are arranged according to a definite pattern, which is nearly symmetric in the eggs with six SCA (Fig. 6), but clearly asymmetric in those with seven (Fig. 7). In both cases there is a wider gap between the SCA on one side.
This description already points to a general similarity between the pattern of SCA and the pattern of follicle cells discussed above. A closer inspection shows that the two patterns are nearly superimposable. In Fig. 6 the average positions of the centres of the follicle cells in 6-cell follicles have been indicated by arrows. (It should be noted that in Fig. 2 the follicle cell patterns have been represented as seen from the animal pole, whereas in the reconstruction of the SCA in Fig. 6 the egg is viewed from the vegetal side; therefore, the follicle cells from 1 to 6 here follow each other in a clockwise direction). It is evident that there is a striking similarity between the two patterns, each of the SCA strictly corresponding to the position of a single follicle cell. In Fig. 7, where eggs with seven SCA are compared with 7-cell follicles, the correspondence between the two is not as good, but in view of the fact that the opposition of four follicle cells on one side against three on the other is matched by the arrangement of the SCA, it seems satisfactory. It should be stressed that the patterns of follicle cells on the one hand, and SCA, on the other, have been established by two independent lines of research, using different methods, and that their great similarity only revealed itself in the final elaboration of the results. Therefore, the two mutually confirm each other, and strongly argue for the reality both of the pattern of follicle cells and of SCA.
A difficulty might seem to arise from the fact that only eggs with six or with seven SCA have been found, whereas many egg follicles in L. peregra have eight or even nine cells. If the pattern of SCA is supposed to be a reflection of the follicle cell pattern, one should therefore expect also to find egg cells with eight or nine SCA.
For the moment the most plausible explanation of this discrepancy is provided by the observation that, while originally the follicle cells are tightly apposed to the oocyte surface, during later growth stages a follicle cavity appears between the two. This has been observed both in L. stagnalis (Ubbels, 1968) and in L. peregra. The formation of the follicle cavity coincides with the beginning of the final stage of rapid growth; in L. peregra at that time about six follicle cells have been formed (Fig. 1). At this stage also the formation of a zona radiai a becomes evident, pointing to a change in the superficial regions of the oocyte. One may assume that the influence of the follicle cells, ‘imprinting’ their pattern on the egg surface, is greatly reduced or entirely abolished when they withdraw from the oocyte. In that case, the mosaic pattern of the egg cortex will not represent the final configuration of the follicle cells, but rather their arrangement at the moment of formation of the follicle cavity. Future research must show whether this supposition is right.
We may conclude that the results of this investigation lend support to the hypothesis that part of the developmental information is transmitted from the parent to the offspring by way of the egg follicle, whose structure is ‘imprinted’ upon the egg during oogenesis. It seems hardly probable that the correspondence between the follicle cell pattern, on the one hand, and the arrangement of the SCA of the oviposited egg, on the other, in both L. stagnalis and L. peregra is merely accidental. It has been shown in L. stagnalis that the polarity and dorso-ventrality of the later embryo arise in conformity to this pattern (Raven, 1967). It is therefore likely that the structure of the egg follicle determines the polarity and dorsoventrality of the egg cell. However, the present investigation seems to show that the asymmetry of the embryo is not determined in a similar way.
SUMMARY
The structure of the egg follicles in dextral and sinistral Limnaea peregra has been studied.
The number of follicle cells surrounding large oocytes varies between 6 and 9.
The follicle cells are not arranged in an arbitrary way, but according to a definite pattern. This pattern is polar and dorsoventral, and either nearly symmetric or asymmetric depending on the cell number.
The patterns of follicles with different cell numbers are connected in a simple way.
The patterns of follicles in genetically dextral and sinistral snails do not mirror each other, but are identical.
In uncleaved eggs of L. peregra, a system of six or seven subcortical accumulations (SCA) is found. In eggs with six SCA their arrangement is identical with that of the cells in 6-celled follicles; in eggs with seven SCA there is a great resemblance to the cell pattern in 7-celled follicles.
It is concluded that the pattern of SCA reflects the arrangement of the follicle cells at the moment when the follicular cavity begins to form.
The results lend support to the hypothesis that the polarity and dorsoventrality of the later embryo are determined by the structure of the egg follicle. The determination of the asymmetry of the embryo takes place in another way.
RÉSUMÉ
La structure des follicules germinatifs des Limnaea peregra dextres et sinistres a été étudiée.
Le nombre des cellules folliculaires entourant les grandes ovocytes varie entre 6 et 9.
Les cellules folliculaires ne sont pas arrangées de façon arbitraire, mais selon un patron distinct. Celui-ci est polaire et dorsoventral, et symétrique ou asymétrique dépendant du nombre des cellules.
Les patrons des follicules dont le nombre des cellules est différent sont liés d’une manière simple.
Les patrons des follicules des animaux dextres et sinistres ne sont pas inverses mais superposables.
Dans les œufs insegmentés de L. peregra il y a un système de 6 ou 7 accumulations souscorticales (SCA). Dans les œufs à 6 SCA, leur arrangement est identique à celui des cellules dans les follicules à 6 cellules; dans les œufs à 7 SCA leur arrangement se ressemble au patron cellulaire des follicules à 7 cellules.
On conclut que l’arrangement des SCA réfléchit le patron des cellules folliculaires au moment oû la cavité folliculaire commence à se former.
Les résultats soutiennent l’hypothèse que la polarité et la dorsoventralité de l’embryon sont déterminées par la structure du follicule germinatif. La détermination de l’asymétrie de l’embryon se fait d’une autre manière.
Acknowledgements
We are indebted to Professor C. H. Waddington for his kindness in providing us with dextral and sinistral L. peregra from a stock kept in the Institute of Animal Genetics in Edinburgh.