1. Growth, morphogenesis and cell movements were studied in Obelia loveni, O. geniculata and Dynamena pumila with the use of time-lapse cinematography, visual observations of vitally stained objects and by histological techniques.

  2. Growth pulsations with the period around 14 min and the amplitude around 15 μm exist in Dynamena pumila and with the period 5–8 min and amplitude up to 5 μm in Obelia loveni. It was demonstrated that the rhythm of growth pulsations does not coincide with the rhythm of periodical contractions of the proximal part of coenosarc.

  3. The distalwards movements of individual cells in the ectoderm of growing stems and hydranth rudiments are described. A considerable variability in the rates of movements of ectodermal cells has been demonstrated.

  4. Different kinds of cell reorientations in developing rudiments are described. As a rule, they precede the alterations of growth directions or of rudiment shapes.

  5. The mechanisms involved in deformations of epithelial layers are discussed.

  6. The possibility of the existence of passive, elastico-plastic structures in the deforming epithelial sheets is suggested.

The problem of co- ordination of cell movements during morphogenesis attracts special attention nowadays. It has been extensively studied in sea-urchin embryos (Gustafson & Wolpert, 1967) and several other species. Recently a number of investigators have concluded that hydroid polypes also are suitable organisms in this respect. The simplicity and geometrical regularity of their shape and the relative uniformity of their cellular structure permits the hope of working out some general principles of the organic formation of shape.

A comprehensive review of Hydrozoa morphogenesis was made by Webster (1971). Many (perhaps, a majority of) studies mentioned there deal with the problems concerning ‘global’ morphogenetic patterns – that is, determination of polarity, general subdivision of a body, etc. More ‘local’ problems concerning the morphogenetic activities of individual cells and their spatial and temporal co- ordination seem as yet to be insufficiently studied, especially in marine Hydrozoa. The task of this paper is to describe these aspects of morphogenesis in some marine Hydrozoa, belonging to subclass Thecaphora.

An important role of cell movements in establishing morphological patterns of Hydrozoa has been stressed by various authors (for review of data on marine Hydrozoa see Berrill, 1961). According to a number of authors (Beloussov, 1961 a, 1963;Hale, 1960, 1964; Crowell, Wyttenbach&Suddith, 1965; Filatcheva, 1966; Campbell, 1967) cell proliferation does not play any direct role in determining the shape or even the dimensions of rudiments. As to the temporal pattern of cell movement, some years ago an interesting phenomenon of growth pulsations in the tips of Hydrozoa outgrowths was described (Beloussov, 1961 b;,Hale, 1964; Wyttenbach, 1968). The most detailed description of the phenomenon was made on Campanularia stolons by Wyttenbach. The following points of his work should be mentioned here : growth pulsations of a stolon tip originate independently from the periodical contractions of the proximal part of the coenosarc; the duration of the pulse cycle is about 6 min at 20 °C and decreases along with the rise of temperature. At constant temperature, cycle duration is much more stable than the amplitude of pulsations. A periodical thinning of the ectodermal layer correlated with the upward shifts of the entodermal layer was also observed.

These conclusions will be later compared with the original data. In the present paper the results of investigations of growth pulsations and of the morphogenetic movements of individual cells are described and some problems concerning the mechanisms of deformations of epithelial sheets are discussed.

Three species of marine Hydrozoa (Thecaphora) have been studied: Obelia loveni, O. geniculata and Dynamena pumila. The samples were collected on the tide lands of the White Sea. The present paper deals only with the development of the vegetal generation.

The following methods of vital observations were employed :

(1) Time-lapse films. Several periods of stem growth and hydranth formation in Obelia loveni and Dynamena pumila were registered on 35 mm film. Exposure intervals ranged from 0-5 min to 2 min. The duration of continuous filming was up to 12 h. A ‘Convas’ camera was used, combined with microscope MBI-3. The microscope objectives were×3 and water immersion×40. The samples were immersed in a 20 ml glass dish, being attached to its bottom by an adhesive tape. The temperature varied from 15 to 20 °C.

(2) Vital staining. Stems and hydranths of Obelia loveni were immersed for 40–60 min in weak solutions of Nile blue sulphate. Then they were transferred to pure sea water and fixed horizontally ; 1–2 h later the staining became granular. The granules were localized in ectodermal cells exclusively. The movement of several (up to four) granules was continuously traced during 2–3 h with the use of a water-immersion objective,×40.

3. For histological purposes samples of colonies were fixed in a Bouin solution and embedded in paraffin wax or paraffin-celloidin ; 8–12 μm sections were prepared. The slides were stained with Heidenhain’s iron haematoxylin.

I. General description of structure and development of vegetal generation of the studied species

A. Some notes on histological structure

The body of a Hydrozoa colony (usually called its coenosarc) consists of two cell layers (ecto- and entoderm), each being mainly composed by myoepithelial cells. There are some important differences in the structure of ecto- and ento-dermal myoepithelial cells. Whereas the entodermal cells are more epithelial-like and are bound together in a single sheet, the ectodermal ones look like the bundles of fine contractile tonofibrils. They are not strictly connected, but are firmly attached to the common surface membrane, enveloping the ectodermal layer. The surface membrane is not to be confused with the chitinous hydrotheca, or perisarc, which is secreted by special glandular ectodermal cells and which is laid down at the external side of the surface membrane. If a part of coenosarc contracts (e.g. when the hydranth ‘neck’ is formed, see below), its perisarc remains fixed and may therefore be used as a stable point of reference.

B. The external character of branching and hydranth development

Obelia loveni and O. geniculata (Fig. 1)

The branching of the colonies is sympodial.1 The branches of both species are bent, in O. geniculata the bending being more pronounced than in O. loveni. The new stems are initiated either just below the growing tip of the maternal stems or on the proximal parts of the colony. In the first case the new (daughter) stems are always formed at the convex side of the maternal ones (Fig. 1 A, d.s.) and grow in the same plane. Just after their initiation the daughter stems grow parallel to the maternal ones (Fig. 1B, d.s.), but later their tips deviate sideways (Fig. 1C, D, d.s.). In the proximal regions new stems are initiated in the axiles of the old ones. The plane of bending of the first ones is in no way correlated with the plane of bending of the latter.

Fig. 1.

Successive stages of development of stem and hydranth in Obelia loveni (pictures from time-lapse film), d.s., Daughter stem; gr1, first series of grooves; gr2, second series of grooves; h.r., hydranth rudiment; n, hydranth ‘neck’; dp, diaphragm; a, side of a hydranth, adjacent to maternal stem; b, side away from maternal stem; ps, perisarc.

Fig. 1.

Successive stages of development of stem and hydranth in Obelia loveni (pictures from time-lapse film), d.s., Daughter stem; gr1, first series of grooves; gr2, second series of grooves; h.r., hydranth rudiment; n, hydranth ‘neck’; dp, diaphragm; a, side of a hydranth, adjacent to maternal stem; b, side away from maternal stem; ps, perisarc.

As stem growth proceeds, several circular grooves are formed one after another at some distance from the tip (Fig. 1 B, C, grf fixed by the chitinous perisarc. In general, about 3–1 such grooves are formed, and then a period of ‘smooth’ growth takes place, being in its turn succeeded by a new period of groove formation (Fig. 1D–G, gr2). Later the part of the stem located distally from the last groove expands and elongates, thus forming a rudiment of a hydranth (Fig. 1E, A.r.). The proximal part of the rudiment is constricted, forming a hydranth’s ‘neck’ (Fig. IF, G, n). A thin cellular plate, a diaphragm, develops proximally to the neck (Fig. 1G, dp). The most expanded part of the rudiment, situated distally to its neck, later splits into a series of columnar tentacle rudiments, whereas its roof transforms to a hypostome. A hydranth becomes slightly asymmetrical in the plane of the stem bending : the side of a hydranth adjacent to the maternal stem (Fig. 1E–G, a) is slightly more convex than the opposite side. At the same time the hydranth is completely symmetrical in the plane perpendicular to that of stem bending.

Dynamena pumila (Fig. 2)

The branching is monopodial.1 The vertical branches are composed of a series of storeys, each including a pair of identical lateral (LR) and a central rudiment (CR) – Fig. 2 A. The LA’s are bent distally (Fig. 2B–D) and later transform to hydranths, whereas the CR remains undifferentiated and grows farther. At first its tip is narrowed, but later it expands and is transformed to a spherical (Fig. 2B) and then to a triangular (Fig. 2D) rudiment. The plane of its expansion completely coincides with that of preceding CR. Thus the whole colony is flattened in a single plane (Fig. 2B1 D1 E1). After expansion each new CR is split by two vertical furrows into three rudiments (Fig. 2E,1,LR1, CR1), the lateral ones being again transformed to hydranths, whereas the central one grows farther. Similar to Obelia hydranths, the LA’s are subdivided into a diaphragm, a constricted proximal part and an expanded distal part (Fig. 2D, dp, n).

Fig. 2.

Successive stages of development of growing tip in Dynamena pumila (pictures from time-lapse film). A–D, Side view; B1–E1, top view. CR, Central rudiment; LR, lateral rudiment. B, C, The opposite phases (extension and contraction) of the same growth pulse, (dp, Diaphragm; n, hydranth neck.)

Fig. 2.

Successive stages of development of growing tip in Dynamena pumila (pictures from time-lapse film). A–D, Side view; B1–E1, top view. CR, Central rudiment; LR, lateral rudiment. B, C, The opposite phases (extension and contraction) of the same growth pulse, (dp, Diaphragm; n, hydranth neck.)

The following general features of the described morphogenetic processes are to be emphasized: in respect to the directions of their growth, the abovementioned rudiments are, as a rule, either asymmetrical (Obelia) or anisotropic (Dynamena). The plane of anisotropy or asymmetry remains constant for a considerable part of colony.

II. The character of growth of Obelia loveni and Dynamena pumila according to time-lapse films

By means of time-lapse cinematography the pulsatory character of the growth, the relations between growth pulsations and coenosarc contractions and between the rates of migration of the tip cells of both layers (ecto- and entoderm) have been revealed.

A. Growth pulsations

The pulsatory character of growth is most obvious in CR of Dynamenapumila (Fig.. 2B, C; Fig. 3). Short (2–4 min) periods of rudiment contractions (Fig. 2C) are alternated with the more prolonged (10–13 min) of its extension (Fig. 2B). Thus the duration of the whole cycle is about 12–15 min, its amplitude about 15–20 μm. The average rate of growth is fairly constant during the whole period of observations. No obvious correspondence has been observed between the phases of tip pulsations and coenosarc contractions (the latter measured by the length of the coenosarc diameter along BB level, Fig. 3). This corresponds to the results obtained on Campanularia stolons (Wyttenbach, 1968).

Fig. 3.

Pulsations of growing tip (a) and coenosarc (b) in Dynamena pumila. 1, Initial; 2, final stage of a studied period. BB indicates the level of the measured coenosarc diameter. The dotted arch at 2 indicates the position of the distal surface of the rudiment at the initial stage 1. Abscissa: time of observation. Ordinate: position of a growing tip (a) and diameter of coenosarc (b).

Fig. 3.

Pulsations of growing tip (a) and coenosarc (b) in Dynamena pumila. 1, Initial; 2, final stage of a studied period. BB indicates the level of the measured coenosarc diameter. The dotted arch at 2 indicates the position of the distal surface of the rudiment at the initial stage 1. Abscissa: time of observation. Ordinate: position of a growing tip (a) and diameter of coenosarc (b).

Similar growth pulses without obvious correlations with coenosarc contractions are observed on LR’s. However, due to the asymmetry of the rudiments, it is difficult to represent these results graphically.

In Obelia loveni the amplitude of tip pulsations is smaller and less constant. However, using water-immersion×40, periods of 5 min pulsations are constantly observed in Obelia stems and those of 8 min in Obelia hydranths. The period of pulsation is much more stable than its amplitude (Fig. 4).

Fig. 4.

Pulsations of growing tip in Obelia loveni. A, B, C, Successive periods of stem growth, separated by 1 h periods. Abscissa: time of observation. Ordinate: position of a growing tip.

Fig. 4.

Pulsations of growing tip in Obelia loveni. A, B, C, Successive periods of stem growth, separated by 1 h periods. Abscissa: time of observation. Ordinate: position of a growing tip.

In developing Obelia hydranths the periodical shape alterations correlated with growth oscillations can be observed. As a rule, in the phase of maximal contraction the outlines of a hydranth roof become smooth, the protuberances of the tentacles and hypostome disappear. On the contrary, during the extension phase the rudiment outlines become more pronounced and the mentioned protuberances become visible again (Fig. 5A, B).

Fig. 5.

Growth pulsations and fluctuations of shape in Obelia loveni hydranth. A, Fluctuations of shape of right side wall of a hydranth; □, contours with most relief. B, The contours of the whole rudiment, corresponding to 1 (solid line) and 2 (broken line) in A. C, Correlation between the fluctuations of a growing tip (1) and contractions of coenosarc (2); Abscissa: time of observation. Ordinate: position of a tip (1) and diameter of coenosarc (2).

Fig. 5.

Growth pulsations and fluctuations of shape in Obelia loveni hydranth. A, Fluctuations of shape of right side wall of a hydranth; □, contours with most relief. B, The contours of the whole rudiment, corresponding to 1 (solid line) and 2 (broken line) in A. C, Correlation between the fluctuations of a growing tip (1) and contractions of coenosarc (2); Abscissa: time of observation. Ordinate: position of a tip (1) and diameter of coenosarc (2).

In Obelia stems as well as in Dynamena the rhythm of growth pulsations does not coincide with the rhythm of coenosarc contractions (Fig. 6B). In hydranth rudiments, however, both rhythms seem almost completely to coincide (Fig. 5 C). The hydranth rudiments behave thus as a unitary contractile system.

Fig. 6.

Different kinds of rate correlations of ecto- and entodermal tip cells in Obelia loveni. A A, BB, contractions of coenosarc diameters at the levels indicated at right. Abscissae : time of observations. Ordinates : positions of the tips of ectodermal (ect.) and entodermal (end.) layers and diameter of coenosarc (AA, BB).

Fig. 6.

Different kinds of rate correlations of ecto- and entodermal tip cells in Obelia loveni. A A, BB, contractions of coenosarc diameters at the levels indicated at right. Abscissae : time of observations. Ordinates : positions of the tips of ectodermal (ect.) and entodermal (end.) layers and diameter of coenosarc (AA, BB).

B. Correlations of migration rates of ecto- and entodermal tip cells in Obelia outgrowths

According to time-lapse data, three different kinds of rate correlations are taking place :

  1. The rates of movement of tip cells in both layers are approximately equal (Fig. 6 A). This situation is mainly typical of the smooth growth period when no grooves appear (a period between the formation of n1 and n2, Fig. 1D).

  2. The rate of movement of ectodermal tip cells exceeds that of the entodermal cells. As a result, the thickness of the ectodermal layer can be approximately doubled (Fig. 6B). This situation is mainly typical of the first period of groove formation (just after the initiation of a stem) – Fig. 1B, n1.

  3. The rate of migration of ectodermal cells is less than that of entodermal cells (Fig. 6C). As a result, the thinning of the ectodermal layer and the more compact arrangement of the tip cells of the two layers take place. The situation is typical of the early stages of hydranth development. Later the reorientation of cells of both layers and the expansion of the rudiment take place (see details below).

The described kinds of rate correlations are much more various than those described by Wyttenbach for Campanularia stolons. Wyttenbach observed only a periodical ‘pushing’ of ectodermal cells by entodermal ones, which led him to conclude the passivity of ectoderm. Meanwhile, in vertical Obelia stems along with the analogous rate correlations (type 3), quite the opposite (type 2) is observed. The latter is incompatible with the assumption of passive ectodermal shift. As a whole, the existence of the different types of rate correlations demonstrates the autonomous migratory activity of cells of each layer.

III. The action of metabolic inhibitors on the shape of the hydranth rudiment in Obelia loveni

Pieces of Obelia colonies, including hydranth rudiments, were immersed in 0·05 M-KCN or 1–2 % dinitrophenol at different times after the beginning of contraction of its proximal part (i.e. neck formation, Fig. 1F). In samples immersed not later than 1 h after the beginning of neck formation, the latter completely disappeared in ; in other words, the proximal parts of the rudiments expanded again (23 cases out of 26). Vital staining and histological examination did not reveal any signs of cell degeneration at the time of expansion. In samples immersed in inhibitor solutions later than 1 h after neck formation, the latter failed to expand even if complete cell degeneration took place. Thus, one may assume the existence of a certain critical point in hydranth morphogenesis after which the contraction of the proximal part is stabilized.

IV. Morphogenetic movements of individual cells

A. Distalwards movements of ectodermal cells according to observations in vivo

Observing O. loveni outgrowths, stained in vivo, the movements of the individual ectodermal cells were traced. The main results of the observations were as follows.

The cells situated at the distal region of growing stem and hydranth rudiment migrate distalwards at a rate approximately equal to the growth rate of the whole stem (around 30 μm/h). At the same time a considerable disparity in migratory rates of individual cells was revealed (Fig. 7). In order to evaluate the degree of rates dispersion and its dependence on the relative positions of the observed cells, a number of proximal (situated between two neighbouring grooves) and of distant (separated by grooves or situated at the opposite sides of the stem) cells were traced. In all, the movement of 48 pairs of proximal stem cells, 24 pairs of distant stem cells, 35 pairs of proximal hydranth cells, 35 pairs of analogous proximal cells of hydranth rudiment and 36 pairs of distant hydranth rudiment cells was traced and the differences of simultaneous 10 min distalwards cell tracks were measured. The cumulative cell rate differences distribution curves were plotted and were compared with the corresponding theoretical curves of normal (Gauss) distribution (Fig. 8). One can see that the dispersion in cell rates for the proximal cells does not differ considerably from that for distant cells. The rate differences distribution curves for stem cells were distinctly subnormal (excess index e = 2·2 ± 0·32), whereas the similar curves for hydranth rudiment cells did not differ sufficiently from the corresponding Gauss curve (e = 0·24 ±0·36). These results demonstrate that the distalwards migration of stem cells is to some extent more co- ordinated than that of hydranth cells. However, in both cases the difference in rates of even proximal cells is great enough to demonstrate the lack of strict connexions between them. That seems to be the most important conclusion from the above data.

Fig. 7.

Trajectories of migration of four ectodermal cells in O. loveni stem. Left, graphics; right, cell arrangement in growing stem. Abscissa: time of observation; ordinates: positions of observed cells.

Fig. 7.

Trajectories of migration of four ectodermal cells in O. loveni stem. Left, graphics; right, cell arrangement in growing stem. Abscissa: time of observation; ordinates: positions of observed cells.

Fig. 8.

Cell rates differences distribution curves (cumulates). Abscissa: differences in rates of simultaneous migration of cell pairs (μm per 10 min). Ordinate : percentage of cell pairs. Solid lines, stem cells; dotted lines, hydranth cells. ○, Distant cells; ×, proximal cells; △, theoretical curves of normal (Gauss) distribution. Right below, a scheme of arrangement of pairs of proximal (ab), and distant (ac, be, cd, bd, ad) cells.

Fig. 8.

Cell rates differences distribution curves (cumulates). Abscissa: differences in rates of simultaneous migration of cell pairs (μm per 10 min). Ordinate : percentage of cell pairs. Solid lines, stem cells; dotted lines, hydranth cells. ○, Distant cells; ×, proximal cells; △, theoretical curves of normal (Gauss) distribution. Right below, a scheme of arrangement of pairs of proximal (ab), and distant (ac, be, cd, bd, ad) cells.

At the more proximal stem regions the cell movement is of a pulsatory character, without a significant distalwards shift. Here the cell shifts seem to be correlated with coenosarc contractions and may be regarded as passive.

B. Cell movements at the growing stem tip according to histological data

The growth of stem tip consists of sliding distal shifts of the central group of cells. In some cases these shifts are relatively rapid (about 90 μm/h) and spraylike (Fig. 9 A). As a rule, however, the rate of the distal shifts of tip cells does not exceed 30 μm/h. The successive phases of the shift are presented on Fig. 9B–D.

Fig. 9.

Different phases of cell shifts in the growing tip of Obelia stem. b.m., Basal membrane; r.b.tn., rupture of basal membrane.

Fig. 9.

Different phases of cell shifts in the growing tip of Obelia stem. b.m., Basal membrane; r.b.tn., rupture of basal membrane.

One can see that after the extension of the ectodermal cells the rupture of basal membrane takes place (Fig. 9, r.b.m)obviously due to the pressure of entodermal cells, which later also shift distalwards.

The sliding shift of the central cell group in respect to lateral ones as well as the alternating character of ecto- and entodermal cell shifts demonstrate the lack of close connexions between the tip cells inside each layer and between both layers.

C. Reorientations and deformations of cells

One of the most interesting features of cell behaviour in Hydrozoa morphogenesis seems to be various reorientations and deformations of cells, regularly correlated with the shape alterations of the whole rudiment and as a rule preceding the latter.

(a) Reorientations in bending rudiments

The orientation of cells in straightly growing rudiments is symmetrical. Often their arrangement at the rudiment tips is similar to that of the onion scales (Fig. 9D). However, in the rudiments which are to be bent soon, cell orientation becomes asymmetrical. In such rudiments the bisector of the angle formed by the axes of cells at the opposite walls deviates considerably from the rudiment axis in the direction of the presumptive bending (Fig. 10A, B). The corresponding angle varied from 6° to 25° (17° on average) for Obelia loveni (6 measurements) and from 9° to 30° for O. geniculata (20° on average, 5 measurements). No cases of negative or zero-deviation occurred.

Fig. 10.

Cell orientation in stems of Obelia geniculata (A) and O. loveni (B) just prior to their bending. AO and BO indicate cell orientation in the opposite stem walls. OK, Bisector of angle AOB. Dotted lines indicate median axes of stems.

Fig. 10.

Cell orientation in stems of Obelia geniculata (A) and O. loveni (B) just prior to their bending. AO and BO indicate cell orientation in the opposite stem walls. OK, Bisector of angle AOB. Dotted lines indicate median axes of stems.

The correlation between cell orientation and the direction of rudiment growth was demonstrated experimentally in D. pumila (for details see Beloussov, 1965). Normally the cell arrangement in the LR’s of D. pumila is clearly asymmetrical, correlated with the presumptive bending of these rudiments. However, if we remove CR, the cell arrangement in the LR’s becomes symmetrical, which corresponds with the vertical direction of their succeeding growth. In CR’s the situation is the opposite. Normally it grows straightly vertical and the cells are arranged symmetrically in its walls (Fig. 11 A). After the removal of LR, however, the cells of the wall, adjacent to the removed rudiment, become more perpendicular to the long axis of the colony. As a result, the whole cell arrangement becomes asymmetrical (Fig. 11B). Such CR’s later grow asymmetrically, bending towards the removed LR. In all cases described above, either normal or experimental, cell reorientation regularly precedes the bending of the whole rudiment.

Fig. 11.

Cell orientation in side walls of CR in Dynamena pumila. A, Normal case; B, after the removal of the right LR.

Fig. 11.

Cell orientation in side walls of CR in Dynamena pumila. A, Normal case; B, after the removal of the right LR.

(b) Reorientations and deformations of cells during the development of hydranth of Obelia loveni and CR of Dynamena pumila

Before the beginning of the expansion of the growing tips in Obelia and Dynamena the orientation of their cells only slightly deviates from the vertical. The lateral cell poles are always situated more distally than the median ones. At the early stages of hydranth (resp. CR) formation the cells of both their layers begin to rotate, tending to orient transversely (Fig. 12). This rotation is especially pronounced in the dis-tolateral parts of the rudiments, which are mostly expanded. At the advanced stages the most complicated cell reorientations take place in the ectoderm (Fig. 13). At the intermediate stages the cells become U-shaped (Figs. 13B, 14A) andin some cases S-shaped (Fig. 13C). Later they become oriented proximally by their lateral ends (Fig. 13E). Therefore, the whole angle of their rotation is about 90° (compare Fig. 13 Aand E). The part of the rudiment composed of such reversed cells later contracts and transforms to its neck. Meanwhile cell orientation in more proximal regions (near the rudiment bases) remains unchanged. As a result, a narrow zone of counter- oriented cells is established near the hydranth (resp. CR) base (Fig. 13E, dp). Later it transforms to a diaphragm.

Fig. 12.

A diagram of cell reorientation at the early stages of development of O. loveni hydranth (left below, initial stage; upper right, final stage of a given period). The arrow on the abscissa indicates the beginning of transverse extension of a rudiment,–×–, Ectodermal cells; –◯– entodermal cells.

Fig. 12.

A diagram of cell reorientation at the early stages of development of O. loveni hydranth (left below, initial stage; upper right, final stage of a given period). The arrow on the abscissa indicates the beginning of transverse extension of a rudiment,–×–, Ectodermal cells; –◯– entodermal cells.

Fig. 13.

Reorientations of ectodermal cells in left side wall of O. loveni hydranth at successive stages of its development, dp, Rudiment of diaphragm.

Fig. 13.

Reorientations of ectodermal cells in left side wall of O. loveni hydranth at successive stages of its development, dp, Rudiment of diaphragm.

One can see from Fig. 13 that during cell orientations the wall of a hydranth rudiment does not considerably alter its shape. At the same time cell reorientations are closely correlated with the presumptive deformations of the sheet. It can be demonstrated geometrically : if we prolong the cell axes (in the case of D. pumila CR, the distal parts of the axes) and plot a curve, perpendicular to all these segments, the curve is obviously similar to the outline of the succeeding stage of the given rudiment (Fig. 14).

Fig. 14.

Three successive stages of development of CR in Dynamena pumila. The line perpendicular to the distal segments of the ectodermal cell axes is approximately similar to the contour of the succeeded stage.

Fig. 14.

Three successive stages of development of CR in Dynamena pumila. The line perpendicular to the distal segments of the ectodermal cell axes is approximately similar to the contour of the succeeded stage.

Specific alterations of cell shape and arrangement take place in the ectoderm of the hydranth rudiment roof in Obelia. The early steps of hydranth formation are characterized by the extensive shortening of the roof ectodermal cells (compare Fig. 9B-D and Fig. 15A, B). Somewhat later they become bow-shaped (Fig. 15A), then they straighten out again (Fig. 15B). At this stage the cell arrangement in the rudiment roof is in many cases more relief than the outlines of the rudiment surface (Fig. 15B1; compare the dotted line which is parallel to the arrangement of cell nuclei with the solid line which is parallel to the rudiment surface). One can see that the line of nuclei arrangement is similar to the dotted line on Fig. 5B, while the outline of the hydranth surface is similar to the solid line of the same figure. Remembering that the two lines reflect different phases of shape pulsations, one may suppose that during the phase of smoothing of the external surface there are individual cells (or their nuclei) which retain the pattern of the ‘relief’ phase, i.e. the pattern of the morphodifferentiation of a hydranth. Therefore the individual cells may be regarded as the active elements of the rudiment, whereas the common surface membrane of a layer seems to be more passive.

Fig. 15.

A, B, Cell arrangement in the roof of a rudiment of O. lovenihydranth at two successive stages of its development. B1, A schematic representation of the contours of hydranth roof and of the arrangement of cell nuclei at stage B.

Fig. 15.

A, B, Cell arrangement in the roof of a rudiment of O. lovenihydranth at two successive stages of its development. B1, A schematic representation of the contours of hydranth roof and of the arrangement of cell nuclei at stage B.

The phenomena described above pose several problems, some of them being narrower, whereas others are more general. The discussion aims rather to outline these problems than to suggest solutions.

1. Pulsatory character of stem and hydranth growth

This phenomenon undoubtedly requires detailed study. Several aspects seem to be of special interest here. First of all, the existence of such pulsations demonstrates a certain synchrony in behaviour (that is, in degree of extension-contraction) of a large group of tip cells. The origin of this synchronization remains completely unknown. A narrower problem is the relation between tip pulsations and the contractions of the proximal regions of coenosarc. Our data reveal no direct correlation between the two, but perhaps some sort of indirect interdependence between these processes nevertheless exists.

The biological role of growth pulsations seems to be clearer : it is the way to ensure a simultaneous pressure of the whole group of tip cells on the passive sheet membranes and thus to deform or rupture the latter. Without synchronization of cell shifts the mechanical resistance of the membranes (first of all, the basal membrane) could not be overcome.

It is to be stressed, however, that cell shifts are synchronous only at the rudiment tips, whereas, according to in vivo observations, the migration of the more proximally situated cells demonstrates a considerable dispersion of rates at every given period.

2. Cell reorientations and mechanisms of deformation of epithelial layers

Our data suggesting an important role of cell reorientations in Hydrozoa morphogenesis are in agreement with those obtained on freshwater Hydra and several other species (Webster, 1971). This author attributes to cell reorientations a leading role in determining the direction of morphogenetic gradients. This interesting problem is, however, beyond the scope of the present paper. The only aspect to be discussed here is the information about the mechanisms of epithelial deformations which may be derived from this phenomenon.

The problem of the nature and distribution of forces which deform a cell layer in normal development nowadays attracts the attention of several investigators. Some of them suppose that the changes in epithelial shape are caused by active expansion or contraction of its surface (Baker & Schroeder, 1967). Another point of view is that the deformation of a sheet is an immediate result of shape alterations of the cells, both processes occurring simultaneously (Gustafson & Wolpert, 1967). For example, the invagination of sea-urchin entoderm these authors regard as a result of spherulization of its cells, leading to the extension of the whole rudiment surface. Changes in the adhesive cell properties are regarded as a cause of alterations of cell shape.

Therefore, according to the first conception, the surface of a sheet behaves like an active structure, while according to the latter one the active role is played by the individual cells. At the same time both conceptions coincide in assuming the simultaneity of the deformations of the individual sheet cells and the sheet as a whole.

In Hydrozoa, however, we see a different picture. As was mentioned above, the regular reorientations of the individual cells take place considerably earlier than the sheet alters its shape. Thus, the deformation of the epithelial sheet consists of two phases: (1) the phase of cell reorientation without a significant deformation of the whole sheet (latent phase) ; (2) the phase of a visible deformation of the sheet due to the shifts of already reoriented cells parallel to their axes. It may well be that two similar phases occur also in a number of other morphogeneses ; for example, the convergence of the dorsal poles of the neural plate cells takes place considerably earlier than the plate starts to evaginate. In this case the problem of epithelial deformation may be separated into two main questions: (a) What is the cause of cell reorientation? (b) What are the mechanisms of cell shifts, succeeding cell reorientations ?

  • This problem may be discussed now only in general terms. So far as cell reorientations determine the regular morphological pattern of the entire rudiment, one must seek for the explanation of the regular spatial distribution of the reorienting forces. This kind of problem belongs to the realm of morphogenetic fields theories, or to its newest modification, a ‘positional information’ concept (Wolpert, 1969). On the other hand, observing cell reorientations (e.g. on Fig. 13), one can see that they need a number of separate forces of various magnitudes and directions, applied to individual cells and even to different parts of a cell. It is hard to imagine such local forces to be originated from a single source, removed from the reorienting cells. On the other hand, it seems to be natural to assume the origin of the reorienting moment in the immediate neighbourhood of a given cell. Therefore one may suppose that the ‘positional information’ which determines cell reorientation is transmitted to each cell by the neighbouring cells. Some years ago a similar model was proposed by one of us (Beloussov, 1968). In this model cell reorientations were regarded as a result of the asymmetrical position of a given cell in respect to the adjacent cells. For example, if a certain cell was surrounded by two others, one of the latter (A) forming a larger angle with the axis of the central cell than the other (B), the first cell tended to rotate towards A-cell. This model gave a formal explanation to cell reorientations described above. The nature of the forces involved was not, however, discussed. Further investigations are needed here.

  • One can see that during the ‘post-reorienting’ cell shifts the shapes of the individual cells are transformed from non-symmetrical (parallelogram-like or bent) to more symmetrical, often rectangular (Fig. 14, compare B and C). If we assume that each cell possesses an intrinsic tendency for symmetrization and at the same time can serve as a support for the shifts of the neighbour cells, the specific pattern of deformation of the whole sheet can be derived. The symmetrization tendency of the individual cells seems to be, in its turn, a natural phenomenon. One can interpret it, for example, as a result of the tendency of the energy of the cell surface to decrease. Undoubtedly, however, this explanation is purely speculative and the problem requires further study.

3. Deforming-resistant structures in epithelial sheet

The possibility of reversing the contraction of the proximal part of the hydranth rudiment by means of metabolic inhibitors may be interpreted as a result of the existence of elastic deforming-resistant structures, which under normal metabolic conditions are overcome by shifting cells, but reveal their tendency to contract after cell depression. If, however, the duration of sheet deformation is sufficiently long, these structures lose their elastic properties and behave as plastic bodies.

Thus, the existence of structures with elasto-plastic properties in epithelial sheets seems to be plausible. It seems reasonable to attribute these properties to the main supporting membranes of the sheet – that is, to the surface and to the basal membranes. The relative morphogenetic passivity of the surface membrane in respect to the individual cells can be derived as well from Fig. 15B and the corresponding comments.

These properties of epithelial membranes can be of morphogenetic significance in preventing small and casual cell shifts (which are taking place from time to time in the proximal regions of Hydrozoa colonies) and in stabilizing the results of massive and vast shifts.

Baker
,
P. C.
&
Schroeder
,
TH. E.
(
1967
).
Cytoplasmic filaments and morphogenetic movements in the amphibian neural tube
.
Devi Biol
.
15
,
432
450
.
Beloussov
,
L. V.
(
1961a
).
Cell proliferation and growth of hydroid polypes (Russian)
.
Z. obsch. biol
.
22
,
281
291
.
Beloussov
,
L. V.
(
1961b
).
Vital observations on cell movements in hydroid Obelia flexuosa (Russian)
.
Dokl. Akad. Nauk USSR
136
,
1490
1493
.
Beloussov
,
L. V.
(
1963
).
The origin of abundant interstitial cells in morphogenesis of Campanularia integra and Campanulina lacerata (Russian)
.
Dokl. Akad. Nauk USSR
150
,
690
693
.
Beloussov
,
L. V.
(
1965
)
Alterations in growth directions as a result of rudiment interactions (Russian)
.
Dokl. Akad. Nauk USSR
160
,
475
478
.
Beloussov
,
L. V.
(
1968
).
Calculations of some cell shifts in hydroid polypes (Russian)
.
Dokl. Vysch. Schkoli (Biol
.)
3
,
7
16
.
Berrill
,
N. J.
(
1961
).
Growth, Development and Pattern
.
New York
:
McGraw Hill
.
Campbell
,
R. D.
(
1967
).
Cell proliferation and morphological patterns in the hydroids Tubularia and Hydractinia
.
J. exp. Zool
.
17
,
607
616
.
Crowell
,
S.
,
Wyttenbach
,
CH.
&
Suddith
,
P.
(
1965
).
Evidence against the concept of growth zones in hydroids
.
Biol. Bull. mar. biol. Lab., Woods Hole
129
,
403
.
Filatcheva
,
L. F.
(
1966
).
Morphogenesis of the lateral outgrowth in Dynamena pumila (Russian)
.
Dokl. Vysch. Schkoli (Biol
.)
1
,
7
10
.
Gustafson
,
T.
&
Wolpert
,
L.
(
1967
).
Cellular movement and contacts in sea urchin morphogenesis
.
Biol. Rev
.
42
,
442
498
.
Hale
,
L. I.
(
1960
).
Contractility and hydroplasmic movements in the hydroid Clytia johnstoni
.
Q. JI microsc. Sci
.
101
,
339
350
.
Hale
,
L. I.
(
1964
).
Cell movements, cell division and growth in the hydroid Clytia johnstoni
.
J. Embryol. exp. Morph
.
12
,
517
538
.
Waddington
,
C. H.
&
Perry
,
M. M.
(
1966
).
A note on the mechanisms of cell deformation in the neural folds of the amphibian
.
Expl Cell Res
.
41
,
691
693
.
Webster
,
G.
(
1971
).
Morphogenesis and pattern formation in hydroids
.
Biol. Rev
.
46
,
1
46
.
Wolpert
,
L.
(
1969
).
Positional information and the spatial pattern of cellular differentiation
.
J. theoret. Biol
.
25
,
1
47
.
Wyttenbach
,
CH. R.
(
1968
).
The dynamics of stolon elongation in the hydroid Campanularia flexuosa
.
J. exp. Zool
.
167
,
333
353
.
1

This type of branching is characterized by a replacement of an old growth point by a new one after the formation of each new outgrowth; every growth point arises proximally to the old one.

1

This type of branching is characterized by a prolonged action of a single growth point which is situated at the distal pole of the growing stem.