Hydroid Stolonal Contractions Mediated by Contractile Vacuoles

In Hydrozoa, and in particular hydroids, a colonially organized polyp generation is the most frequent or even the only generation to develop. The development of a colony starts with the metamorphosis of an actinula or planula larva into a primary polyp, which grows a primary stolon. This stolon can grow and bifurcate in a variety of ways, resulting in different types of two-and threedimensional colony growth forms. The morphological subunits of developing colonies, however, share a common organizational plan (see Fig. 1A). Polyps, which serve as feeding organs, are connected to one another by tubular extensions of a common gastrovascular cavity to form a stolonal vasculature. The vasculature bathes tissues in a common gastrovascular fluid, the movement of which is presumed to be driven by some combination of polyp contractions, stolon contractions and endodermal ciliation (Hauenschild, 1954; Crowell, 1974; Müller, 1974). Colonies grow by elongation of stolon tips, by induction of new tips and by vegetative iteration of new polyps along these stolons (Müller et al. 1987).

Although the importance of gastrovascular flow for hydroid growth and development is unquestioned (Hauschka, 1944; Berrill, 1949; Hauenschild, 1954; Hale, 1960; Fulton, 1963; Karlson and Marferin, 1984; Marferin, 1985; Müller et al. 1987), the contribution of stolonal contractions to these processes remains unclear. Three views are evident in the literature. Berrill (1949), Hale (1960, 1964), Fulton (1963), Hudson (1965) and Donaldson (1974) have suggested that gastrovascular flow is driven by specialized contractile regions that occur in some thecate hydroids just proximal to stolon tips, whereas Hauenschild (1954), Crowell (1974) and Müller (1974, 1975) have suggested that a combination of polyp contractions, stolonal contractions and endodermal ciliation drive gastrovascular flow. Marferin (1985), however, regards stolonal contractions to be an incidental by-product of gastrovascular flow driven by polyp contractions. Of these divergent views, only the relationship between gastrovascular flow and endodermal ciliation has been determined. Experiments performed by Berrill (1949) and Hale (1964) have shown that gastrovascular flow ceases under conditions in which stolonal and polyp contractions are blocked, but endodermal ciliation remains unaltered. Our relative lack of understanding of the forces governing gastrovascular flow is unfortunate in that patterns of gastrovascular flow within colonies are highly stereotyped in both growing (Karlson and Marferin, 1984; Marferin, 1985) and regenerating (Huxley and deBeer, 1923; Hammett and Padis, 1935; Hale, 1964) animals, suggesting the existence of sophisticated mechanisms governing its distribution.

Using the athecate hydroid Hydractinia symbiolongicarpus (Buss and Yund, 1989), we (i) quantify stolon contraction rates under various conditions of feeding, starvation and temperature; (ii) show by experimental manipulation that stolon contractions occur at normal frequencies in the absence of polyps, raising the question of the origin of stolonal contractions; and (iii) present micrographic studies that reveal the presence of a contractile vacuole in the stolon endoderm whose activity correlates with the observed pattern of stolonal contractions.

Clonal propagations of two Hydractinia symbiolongicarpus colonies were used for all experiments. Both clones were highly stoloniferous (‘net type’, Hauenschild, 1954). Clone a (male) was a field-collected colony from Long Island Sound (Guilford, CT); while clone β (female) was derived from a mating of two fieldcollected colonies. Stock colonies were maintained on hermit crab shells in 40-1 recirculating aquaria at 16°C and fed 3-to 4-day-old nauplii of Artemia salina daily. Clonal propagations were established by standard methods (McFadden et al. 1984). Clonal explants were grown on 22mmx40mm cover slips for video microscopy and actin staining. Explants for transmission electron microscopy (TEM) were grown in Falcon Petri dishes (25 mm). All clonal explants were maintained at 21-23 °C and fed every 2 days with 4-to 5-day-old nauplii reared at 16 °C. All experimental colonies were in a non-reproductive state and bearing feeding polyps (hereafter referred to as polyps) as the only type of hydranth. These colonies were kept under artificial daylight conditions of L:D 16h:8h between experiments and under constant light conditions (L:D 24h:0h) throughout the experiments. In addition, colonies of the athecate hydroid Eleutheria dichotoma Quatrefagos, were cultured in a similar manner (Hauenschild, 1956) for the purpose of comparative TEM studies on the ultrastructure of the stolons.

Colonies on cover slips were mounted in a silicon rubber-coated aluminum chamber (110 mm x 40 mm x 15 mm) with a circular hole (16 mm diameter) in the bottom, allowing the microscope objectives to contact the cover slips from underneath the colony. The animal chamber was thermoregulated by using a constant-temperature stage (Rainin, Massachusetts, model 5000 digital controller, model 5002 slide holder), allowing temperature constancy to better than ±1°C. Stolons were videotaped using a Zeiss Axiovert 35 microscope either in bright field (2.5x Plan-Neofluar) or differential interference contrast (DIC) (20x and 40x Achroplane), with or without a 1.6x magnifier. The microscope was equipped with a phototube interfaced to a Dage MTI70 series camera, Mitsubishi time-lapse video cassette recorder HS-480U, Mitsubishi video copy processor P61U, and a high-resolution Panasonic WV5410 black and white monitor. For measuring the time course of changes in stolon canal diameter during contraction-expansion cycles, the video recorder was interfaced with a microcomputer (IBM-compatible PC-AT) equipped with the Bioscan Optimas image-analysis software. 35 mm micrographs were obtained using a Nikon F3 HP camera body and Kodak T-MAX professional 100 ASA black and white film.

The effect of temperature on the rate of stolon contraction was quantified at two different sites, SO (i.e. a main stolon bearing a feeding polyp) and SI (i.e. a side branch without a polyp), within intact colonies, as shown in Fig. IB. SO and SI stolons of clones α-and β were videotaped at 60 x time lapse while temperature was raised continuously from 5 to 31 °C over a period of 12 h. The number of stolon contractions per 30 min interval was counted and plotted against the actual temperature at the beginning of the interval. The frequency of stolonal contraction was measured a day before and a day after the experiment at 18 °C for each experimental colony. Only those colonies that showed no significant effect of experimental manipulation on contraction rates were utilized in the analysis. Different linear and non-linear regression models were tested to fit the data using SAS and BMDP, and a decision on the best fit was based on the deviation values for the residual sums of squares (RSS) (Schierwater, 1989).

The relationship between stolon contraction, polyp contraction and the direction of gastrovascular flow was investigated by simultaneous monitoring of all three variables in an intact growing colony. As above, two different sites (SO, SI) in each of the two clones (α, β) were observed. Stolons were videotaped at 60 x time lapse and at about 200x final magnification on the screen. At 22°C each colony was starved for 24 h, monitored for 24 h, then fed and monitored again for a further 24h. Using a simultaneous on-screen monitoring of time, tapes were analyzed to obtain the number of stolonal contractions, the number of changes in the direction of the gastrovascular flow and the number of contractions of the polyp for hourly intervals. Nonparametric Kendall’s tau correlations were calculated for all the listed variables using BMDP. It is important to emphasize that these measures are only capable of revealing correlations between local events. The field of view was limited to a total area of approximately 1.5 mm² of the colony appearing on the screen and the influence of patterns of polyp or stolonal contractions outside this limited region of the colony will not be revealed by these methods.

To determine whether stolon contractions persist in the absence of direct vascular connection to polyps, we severed the connections between a segment of stolon and the colony. Isolated segments of stolons, ranging in length from 200 to 500μm, were produced in four regions of colonies of clone α: SO, SI, S2 and Sm (see Fig. 1). An additional eight stolons, four SO and four SI, of clone βwere tested in the same way. Stolon contraction rates were quantified by video microscopy as described above for a 2 h period prior to surgery and a 10 h period after surgery. Stolon contraction rates before and after surgery were compared using a Mann-Whitney U-test (two-tailed).

Stolons of both clones, α and β, were examined using TEM to determine whether characteristic intracellular changes accompanied stolonal contractions. Samples were observed in each of three possible states: fully contracted, fully expanded and in transition between contracted and expanded. Stolons of 500 μm length were severed from the mother colony, microvideotaped for 1 h, and fixed in glutaraldehyde as soon as the different contraction stages were found simultaneously along the stolon. Photographs of the stolon were taken at the moment of fixation and subsequently used as a guide for cross sections analysed by TEM. Colonies were fixed for 1 h in 4 % glutaraldehyde (prepared in artificial sea water ‘Instant Ocean’, pH7.4), washed in Instant Ocean, and post-fixed for lh in 1% osmium tetroxide. Colonies were then dehydrated in a graded series of ethanol washes and embedded in Spurr’s epoxy resin. Semi-thin sections (1 μm) were cut with a Sorvall MT-2 ultramicrotome, mounted on glass slides, and stained with Toluidine Blue. Ultrathin sections from silver to gray interference color were cut with a diamond knife, stained with uranyl acetate and lead citrate, and viewed on either a Philips 300 or a Zeiss EM-10 TEM. For strictly qualitative comparisons, stolons from Eleutheria dichotoma were also examined by TEM.

The relative contributions of changes in the volume of ectodermal and endodermal vacuoles to a stolonal contraction were estimated from electron micrographs. The point-counting method (lcmxlcm grid, final magnification 6700 x) was used to estimate the fractional vacuolar volume (ratio of vacuole volume to total volume) of both the endoderm and ectoderm. The same grid was used to estimate cytoplasmic area when determining the density of zipper-like organelles (ZLOs) (see Results, final magnification 50000x).

Rhodamine-Phalloidin was utilized to visualize F-actin in intact stolons (Faulstich et al. 1988). Explants were established on cover slips and grown for a period of 3-4 weeks, by which time they had reached a size of 5-10 feeding polyps with 4-5 branched peripheral stolons. Whole colonies were fixed for 5 min in 2 % paraformaldehyde in phosphate-buffered saline (PBS), rinsed once in PBS, and quenched for 5 min to 50 mmol P¹ NH₄C1, and then incubated in the dark for 5 min with Rhodamine-Phalloidin (Molecular Probes, Eugene, Oregon, no. R415; 1:50 in 1 % Triton X-100 in PBS). The colonies were then washed three times with PBS and mounted with 100 mmol I^-1 n-propyl gallate in 50 % glycerol-PBS, to prevent bleaching. The material was viewed at once, using a Zeiss microscope equipped with a 63x Planapo oil-immersion objective, for both phase and epifluorescence viewing. Micrographs were made with an Olympus OM-2 camera body using Kodak T-MAX professional 400 ASA black and white film, exposed and developed for 1600 ASA.

A complete stolon contraction cycle consists of the contraction phase, during which the canal closes, and an expansion phase, during which the canal reopens (see Fig. 2). In Fig. 3A the time course for a series of five consecutive complete contraction-expansion cycles is shown for a temperature of 25 °C. The time for the stolon contraction phase (mean ±1S.E., 54.7±4.68S) was significantly longer than for the expansion phase (39.8±4.45s) (P<0.05, Wilcoxon, two-tailed). The uneven distribution of the speed at which the canal is closed and reopened is shown in Fig. 3B. This limit cycle uses the same data as the stolon contractionexpansion periodicity in Fig. 3A.

The rate of stolon contraction (i.e. the number of complete contractionexpansion cycles per hour) bears a sigmoid relationship to temperature for all stolons tested. The pooled values for the different clones (ar and and positions within colonies (SO and SI) are shown in Fig. 4. A sigmoid regression curve was found to display a more than 13 % better fit to the data than any linear or simple exponential regression model. The means ±S.D. for the three estimated variables of the best fitting sigmoid regression given in Fig. 4 are 57.4±5.09, 1.3±0.53, 0.178±0.024; RSS=388, d.f.=75. Two aspects of this relationship are noteworthy. First, the range over which stolon contraction rates are most sensitive to temperature (15-25 °C) is within the range of summertime temperatures experienced by these organisms (Buss and Yund, 1988), whereas the extremes of temperature lie outside the range of values naturally encountered by growing colonies. Second, this function is characteristic of that obtained for active, temperature-dependent metabolic processes (Precht at al. 1973; Cossins and Bowler, 1987). The average Q₁₀ value in the temperature range between 15 and 25°C is 2.7 (calculated from the regression). It should be noted that the absolute values for the contraction frequencies might be underestimates, because of the possibility of a delayed physiological response to temperature changes. The sigmoid shape of the regression and the slope (Q₁₀ values), however, would not be affected.

A correlation matrix of stolon contraction rates, polyp contraction rates and changes in the direction of gastrovascular flow is presented in Table 1 for both starved and newly fed colonies. A correlation coefficient of close to 1 would be expected if local polyp contractions alone drive the gastrovascular flow and/or stolon contraction, or if only stolon contractions drive the local gastrovascular flow, or vice versa. Although significant positive correlations were obtained for different variables, no simple 1:1 relationship existed between any two variables. Thus, the findings suggest that the stolonal contractions are not a simple function of local gastrovascular flow or polyp contractions.

Positive correlations between these variables are most frequently observed in starved colonies (Table 1). Feeding often led to obstruction of stolonal canals at different regions, where dense food materials blocked the gastrovascular flow. Clogged canals required 1-8 h to reopen. Of particular interest is the observation that stolon contractions continued in areas between two obstructions. This suggested that stolon contractions might persist in the absence of either polyp contractions or prevailing gastrovascular flow.

The contraction rates of isolated stolons of all four types tested did not differ significantly from the rates observed prior to their isolation (P>0.05, 17-test; Fig. 5). In 9 of 12 such experiments, the isolated stolons grew back and fused to exactly the same point on the mother colony from which they had originally been separated. After fusion to the mother colony, the stolon contraction rate did not differ significantly from that measured for the stolon when isolated (P>0.05, U-test, Fig. 5). These findings suggested the existence of a mechanism for the control of stolonal contraction residing in the stolon itself.

The velocity of the gastrovascular flow in isolated stolons, however, was significantly lower than that observed in intact colonies. In another context we estimated the highest velocities of moving particles in the gastrovascular flow before and after separation of an SI stolon from the colony. Based on the 10 highest values out of a total of 27 measurements, both before and after separation, the velocity decreased by a factor of six in an isolated stolon (35.7±8.03 ;nn s⁻¹ in an intact SI stolon compared to 6.0±1.03μms⁻¹ in the same isolated stolon).

Measurements of the diameter of stolons, the lumen of the canal and the thickness of both the endoderm and the ectoderm were obtained from electron micrographs. The correlation matrix between these variables measured over a single contraction cycle is reported in Table 2. The only significant correlation was a negative relationship between canal diameter and endoderm thickness, suggesting that the opening and closing of canals is associated with events unique to the endoderm (Fig. 6A,B, Table 2). Correlation coefficients relating the canal diameter to the thickness of the ectoderm and the stolon diameter were not significant, suggesting that although changes in the latter two occur they are not of primary importance for the stolon contraction-expansion cycles.

Examination of cross sections in TEM from stolons in all stages of contraction revealed specialized, ‘zipper-like’ organelles (ZLOs). These organelles were found exclusively in apical portions of the endodermal cells, i.e. lining the lumen walls and the walls of the medial vacuoles (Fig. 7A). The ZLOs constituted approximately 2.5-3 % of the endodermal cytoplasm in these regions. At no time were ZLOs observed in basal portions of the endoderm or at any site in the ectoderm. Internally, the organelles consist of two parallel, electron-dense filaments (core) with numerous side arms radiating outwards towards the limiting membrane (Fig. 7B). No electron density was observed between the core filaments. The membrane-to-membrane separation was only 47 nm and the largest continuous ZLO was approximately 0.5 pm in length. The three-dimensional structure of the ZLO is unclear. Occasionally, circular cross sections were visible (Fig. 7A), suggesting a tubular structure. However, such cross sections were relatively rare and this suggests that the ZLO may exist in the form of large, flat sheets.

The appearance of ZLOs differed between stolon contraction stages. In cross sections of stolons fixed while undergoing the transition from the expanded to contracted state, vacuolized ZLOs were found. Here the central core filaments of ZLOs separated (Fig. 7C). Tangential sections through the walls of a vacuole reveal a lengthening and pairing of the side-arm filaments (Fig. 7D). In fully expanded and fully contracted stolons, the vacuolized ZLOs are largely absent. Endodermal cells of contracted stolons are characterized by extensive vacuolization and remnants of ZLOs may occasionally be seen (Figs 6B, 8B), whereas expanded stolons are characterized by few vacuoles in the endoderm (Fig. 6A).

The differences between cross sections taken at differing points in the stolonal expansion and contraction cycle are summarized in Table 3. Compact ZLOs with closed zippers are present in all stages of stolonal contraction. In contrast, vacuolized ZLOs occur at high frequency only in stolonal stages of transition between expanded and contracted. A marked increase in endodermal cell vacuolization accompanies the transition stage and reaches a maximum during the stolonal contraction stage. Within different cross sections, the degree of ectodermal vacuolization was variable. However, no net change in ectodermal vacuolization accompanied the contraction cycle. Thus, differences in the appearance of ZLOs in the endoderm are correlated with vacuolization of the endoderm (Table 3), which is in turn correlated with the opening and closing of the gastrovascular canal (Table 2). ZLOs of an obviously identical ultrastructure were found also only in the stolonal endoderm of Eleutheria dichotoma.

Although epithelio-muscular cells have not previously been noted in hydroid stolons, Rhodamine-Phalloidin staining of intact stolons was performed to exclude the possibility of their occurrence in the stolons of Hydractinia. Rhodamine-Phalloidin bound to sites, i.e. F-actin, in all stolons and polyps of a colony. Binding patterns in polyps (not shown) showed extensive muscular elements, but no such structures were evident in stolons. In the stolons only the endoderm stained intensely. Diffuse actin staining was apparent almost throughout the expanded stolons (Fig. 9A,B), whereas staining was concentrated around the central canal, now closed by endoderm cells, in contracted stolons of the same colony (Fig. 9C,D). The pattern of actin-staining in the endoderm was consistent with the known distribution of ZLOs in the apical endoderm, and it is possible that actin is associated with ZLO function. A decision on whether actin might be part of the membranes of the ZLOs, however, must await immunocytochemistry studies.

The ZLO identified in transmission electron micrographs behaves in a fashion expected of a contractile vacuole. While TEM observations are necessarily static, the contractile nature of ZLOs may be inferred from differences in ZLO organization in the sequence of cross sections derived from stolons at three different stages in the dynamic process of stolonal contraction. In the expanded state, the stolon endoderm is characterized by few vacuoles and by no ZLOs of the vacuolized morphology. The appearance of the vacuolized ZLOs is coincident with the transition from stolonal expansion to contraction and an increase in the extent of endodermal vacuolization that occurs at this time (Tables 2, 3). Finally, in the contracted stolon state, endodermal vacuoles are large and fill most of the lumen of the endoderm cells and vacuolized ZLOs are almost absent (Tables 2,3). These observations provide compelling evidence that the ZLO is a contractile vacuole.

This is the first report of such an organelle from a cnidarian. Amongst the metazoans, only sponges had previously been known to display contractile vacuoles (Gatenby et al. 1955; Brauer, 1975; Simpson, 1984). Contractile organelles are common in protists (Kitching, 1956; Patterson, 1980), where, as in sponges, they occur most frequently in freshwater forms. In these organisms, contractile vacuoles are presumed to function as intracellular pumps acting to maintain cell volume (Patterson, 1980). The ZLOs of Hydractinia, however, act in a different fashion. Rather than expelling water from intracellular stores to the external environment, these organelles act to absorb and release gastrovascular fluid from a common vascular lumen to which they show, at least temporarily, open connections. The ultrastructure of the contractile vacuole of Hydractinia and Eleutheria does not bear obvious structural similarity to the contractile vacuoles of protists and sponges (Patterson, 1980; Akbarieh and Couillard, 1988), and any inference as to evolutionary origins would be premature.

The function of these organelles may be threefold. The gastrovascular fluid is rich in solutes, providing for both respiration and nourishment. We speculate that the contractile vacuoles of Hydractinia act to facilitate both gas exchange and uptake of dissolved organic material from the gastric vasculature. The so-called ‘rod-shaped’ or ‘cytoplasmic’ vesicles in Hydra (Burnett, 1973) are probably the same organelles as the ZLOs described here. These vesicles have been described as phagocytotic, and as adding reversibly to the apical surface membrane of the endoderm (gastrodermis) and also as fusing to larger ‘digestive’ vacuoles (Burnett, 1973). These findings nicely complement our interpretation of ZLOs playing an important role in feeding, i.e. in the uptake and release of gastrovascular fluid from and into the stolonal canals. The question of whether the ZLOs and all larger endodermal vacuoles might be different members of the same population of vesicles cannot be answered at present. However, this study shows that ZLOs probably fulfill two additional functions, independently of their role in feeding. A second function is evident from the observation that ZLO vacuolization acts to close the canal lumen and, hence, to block local gastrovascular flow. Thus, contractile-vacuole-mediated stolon contractions may serve to regulate the spatial and temporal distribution of gastrovascular flow within a colony. Such a mechanism is of particular importance in a vascular system in which a single canal is utilized for both afferent and efferent flow. A third function is evident from our observations of isolated stolons. It is clear from these experiments that the activity of contractile vacuoles is sufficient to initiate and maintain limited gastrovascular flow.

Although stolonal contractions are capable of supporting gastrovascular flow, they are unlikely to provide the main force driving gastrovascular flow in intact colonies, as shown by our observation that the velocity of gastrovascular flow in the absence of polyps is substantially reduced relative to that measured in an intact colony. Karlson and Marferin (1984) have shown that gastrovascular flow can be strongly influenced by the cumulative effect of polyp contractions at sites in the colony distant from the point observed. Under these conditions, only weak correlations may be found between local polyp contraction and local gastrovascular flow (Table 1). If the main gastrovascular flow is propelled by polyp contraction (Karlson and Marferin, 1984; Marferin, 1985), then stolons would generally stay under high gastrovascular fluid pressure generated by distant polyps. This pressure might support a passive inflow into the contractile vacuoles (diastole) during stolonal contraction. As gastrovascular flow is diverted to other regions of the colony and local pressure drops, the contractile vacuole probably contracts (systole) and releases gastrovascular fluid into the gastric system to generate stolonal expansion. It seems likely that forces due to the elasticity of the epithelia and the mesoglea relative to the antagonistic, inelastic periderm also play a role in stolon expansion. The time courses for stolon expansion and contraction phases, as well as Hale’s (1960) observation that the resting state of the stolon in an anesthetized colony is the contracted state, suggest that stolon expansion, i.e. emptying of the contractile vacuoles, is the active phase. Thus, contractile vacuoles may not serve as important forces driving gastrovascular flow, but may play an important role in directing gastrovascular flow by opening and blocking stolonal canals reversibly in response to local pressure differentials (Fig. 10).

Contractile-vacuole-mediated stolon contractions may also bear on the mechanisms of stolon formation in growing colonies. It has been demonstrated that stolon tips can be evoked by mechanical pressure in regions of the colony competent to form stolons (Plickert, 1980; Müller and Plickert, 1982; Müller et al. 1987), as well as by chemical and electrical signals (Müller and El-Shershaby, 1981; Müller et al. 1987). The pattern of stolon contractions in the area of a branching event can, in principle, provide a temporary increase in gastrovascular pressure to drive out the new stolon, by regulating the diameter of canal openings or by regulating the timing of canal opening. The tip region of an outgrowing stolon is a blind end, and hence the diameter of the opening of a canal leading from a main stolon to a tip should determine the hydrodynamic pressure on the tip and, in turn, influence the extent of forward deformation of the tip (Wyttenbach, 1973; Beloussov et al. 1989). Given the lack of nerve cells in stolons, such an effect might be coordinated by a fast epithelial conducting system (Josephson, 1961; Mackie, 1973; Stokes, 1974a,b) acting to control stolon contraction patterns. An epithelial conducting system has been suggested to be present in hydroid stolons on the basis of ultrastructural evidence (Overton, 1963). The role of stolon contractions in mediating stolon tip formation bears further investigation.

Our observations confirm and extend several observations made by prior investigators. As both Berrill (1949) and Hale (1960) independently observed for Obelia commissularis and Clytia johnstoni, respectively, we find that stolon contractions and gastrovascular flow are sustained in the absence of polyps (Fig. 4). Further, Hale (1960) had observed that stolonal contractions are accompanied by large changes in the volume of endodermal cells without comparable changes in the dimensions of the ectoderm (Table 2). In particular, Berrill has noted that in Obelia specialized regions of the stolon are characterized by extensive endodermal vacuolization. Hudson (1965) subsequently showed that these vacuolized regions are contractile in nature. The correspondence between our findings and the observations made on different hydroids suggests that the contractile vacuoles identified in Hydractinia may be characteristic of the vascular system of hydroids.

Finally, our results bear directly on the study of growing tips. Growing tips are specialized regions of stolons that are said to ‘pulsate’ (Wyttenbach, 1968, 1973, 1974; Crowell, 1974; Beloussov et al. 1989), and changes in the shape of the tip region have been attributed to periodic changes in the extent of vacuolization of the tip ectoderm (Beloussov et al. 1989). ZLOs, however, occur in the tip endoderm (and not in the ectoderm) in Hydractinia and Eleutheria, suggesting that changes in the endoderm may also play a role in tip growth. Since the endoderm of the stolonal tip is also highly vacuolated (see Fig. 5 in Beloussov et al. 1989) and since Donaldson (1974) has shown that tip pulsations are blocked by cytochalasin B, it seems likely that the extent of tip elongation is a joint consequence of ectodermal and endodermal vacuolization.

The following predictions can be made if the suggested model for the function of stolonal contractions presented here is correct. (1) If the contractilevacuole-mediated stolon contractions depend upon an active metabolic process, a shortage of ATP (induced experimentally by inhibitors of mitochondrial functions, Blackstone, 1990, or naturally by low oxygen concentrations in the environment) should slow stolon expansion. (2) If nourishment is an important function of stolonal contractions, colony growth rates should follow a sigmoidal temperature dependency similar to that of stolon contraction frequencies. (3) Fast-growing stolons should show high contraction frequencies and/or large contraction amplitudes. (4) SO stolons, which are directly connected to a polyp, should have faster growth rates than SI and S2 stolons. (5) Poorly nourished regions in a larger colony should have relatively more closed stolons, favouring growth in the better nourished regions; i.e. stolon contractions are affected by a positive feedback mechanism. Those stolons growing fast and contracting frequently will be nourished at the expense of slower-growing stolons; consequently, regions of clearly different growth rates should be found at any given time within one colony. These predictions will be tested and used for experimental manipulations in further studies.

Supported by DFG (III-02-Schi-277/l-l), NSF (BSR-88-05961, DCB-90-18003, OCE-90-18396) and ONR (N 00014-89-J-3046). We thank L. Beloussov, N. Blackstone, D. Bridge, C. Cunningham, M. Dick, C. Hauenschild, H. Hadrys, M. Max and M. Peterson for comments on the manuscript and J. Taschner for technical assistance. This paper is dedicated to Professor Carl Hauenschild on the occasion of his 65th birthday and retirement as Professor Emeritus from the chair of zoology at Braunschweig University.