Introduction

The present work was commenced some six years ago by one of the authors (P.M.R.) as a follow-up study arising from an hypothesis (Ralph, 1956) that the temperature range of the three major water masses round the New Zealand coast affected the growth habit of Obelia geniculata. In the subtropical water mass of northern New Zealand, O. geniculata has short stems (5-6 mm) that lack branches, while in the intermediate zone of mixed waters the stems are longer (10-20 mm) and 8%-50% are branched. In the cold waters of the subantarctic zone the colonies are tall (up to 40 mm) and 80%-100% of the stems have branches.

The temperature in Wellington Harbour ranges from approximately 9°C in winter to approximately 20°C in summer. This range seemed sufficiently wide to further test the hypothesis that temperature was a major factor in the growth form attained by the erect stem. The first series of samples for the experiment were collected from Kau Bay (Fig. 1) in September 1961. Sampling continued through 1962 and terminatd in July 1963 as the senior author went on overseas leave. A partial assessment of the data made at this time was of sufficient interest to encourage further sampling if another opportunity arose for collection—particularly for collection in the autumn and winter seasons—as the records for this seasonal range were incomplete. Also, page 2as the original hypothesis formulated that the number of branches increased with decreasing temperature, it was very desirable to have as many samples from this temperature range as possible. The opportunity came to obtain more autumn-winter samples in 1967 and was undertaken by the junior author (H.G.T.) from March to July.

Hammett & Hammett (1945) recorded seasonal changes in colony composition for Obelia geniculata from Provincetown Harbour, Massachusetts, latitude 42.1°N. This is almost the same latitude in the northern hemisphere as the present study area is south of the Equator, namely, 41.2°S. The present study differs however from that of Hammett & Hammett in several aspects.

Firstly, the habitat of the hydroid colonies. The substrate in Wellington Harbour for Obelia is the broad, often longitudinally crinkled lamina of the southern kelp Macrocystis pyrifera (L.). The stolons are most frequently found in the grooves of the lamina, and the erect stems grow singly in rows along the stolons, Many of the laminae float at, or just below the surface of the water. Rarely is O. geniculata found growing on substrates other than Macrocystis pyrifera in New Zealand waters. In Provincetown Harbour, growing stems of O. geniculata were collected either from the shells of barnacles, or, when this habitat was depleted, from Fucus. The colonies grew in dense clusters on the barnacles, but on the Fucus singly "like trees in a grove".

Secondly, analysis of the Wellington Harbour material was made from formalin-preserved random samples. Hammett & Hamett selected living colonies (? the equivalent of our erect stems), "for desirable qualities"; this selection commenced at the time of collection and subsequently through to the setting up of the colonies for growth study in laboratory vessels. Sampling in Wellingon Harbour covered a full seasonal range for two and a half years, and a short "cold-water", 5 month, autumn-winter season. Sampling in Provincetown Harbour covered the "warm-water", 6 month, spring-summer season for seven consecutive years.

Thirdly, Hammett & Hammett regard endogenous chemical factors as solely responsible for seasonal changes in growth in O. geniculata. Our conclusion is that both endogenous and exogenous factors influence growth in this hydroid. The present study gives evidence that assessment of data for part of the seasonal range leads to a biased conclusion. Because of this conclusion, we have described the results of our short-term, autumn through winter sampling before that of the longer term annual data.

Material and Methods

(i) Collection and associated techniques

Colonies of Obelia geniculata were collected from two localities, namely Kau Bay (a weekly sample from September 1961 to July 1963) and from Point Haswell (a biweekly (= twice per week) sample from March to July 1967). The 1961-1963 samples, 48 in all, were subsidiary to a wider programme of harbour plankton collection (Fig. 1). The Obelia medusae from these collections were described by Wear (1965). The data given here for medusae are based on that of Wear.

Kau Bay is a relatively sheltered, wide, horse-shoe shaped bay, with the opening facing north-east. A narrow shingle beach is exposed at low tide. The kelp, Macrocystis pyrifera (L), on which the Obelia grows, occurs commonly along the rocky shoreline in Wellington Harbour at depths of 0.5 to 1.5 fathoms (Fig. 1). At Point Haswell a causeway has been built across the rocky foreshore to the small lighthouse beacon at the seaward extremity.

page 3

Fig. 1 Location map to show areas sampled in Wellington Harbour. (After Wear, 1966).

page 4

During the 1961-1963 period the Obelia was collected by a grapple thrown at random into the belt of kelp, from the deck of the Department's 16-foot launch. The surface temperature was recorded from water collected in a dip-bucket, using a centigrade thermometer graded to 0.2°C. The water, tide, and general weather conditions were also noted. One lamina, occasionally two, of Macrocystis was placed in a plastic one gallon bucket of fresh sea water for transportation to the laboratory. A random segment of approximately 2 cm square was cut from the lamina. The erect stems of Obelia were killed and fixed by placing the cut sample in 250 ml of sea water, and adding 40% formalin drop by drop from a pipette until the concentration of the formol-saline reached approximately five percent. The range of stem height; the number of feeding polyps per stem; the number of stems with terminal buds; the number of branches per stem; the number of gonothecae; the range of internode length, and the range of the distance apart of the stems on the hydrorhiza were recorded for 100 stems from each side of the lamina.

The colonies of Obelia from Point Haswell were collected from a Macrocystis lamina situated about the mid-portion of the stipe. Laminae below this area were usually damaged. Those near the growing point showed little or no colonization by Obelia geniculata. A lamina was cut from the stipe using a three-pronged wire rake attached to a long pole, and transported to the laboratory as in the previous collection series in a clean, plastic bucket of sea water. The sample was cut, killed, fixed and preserved in the manner noted above.

The numbers of Obelia medusae in the plankton catches from Kau Bay were estimated as described by Wear (1965, p. 2), and converted to a percentage of the total volume of the catch.

Remarks

On several occasions weather, and other factors affecting the running of the launch made collection impracticable in the Kau Bay area. Another feature of the original methodology that seemed unsatisfactory was collection by grapple from a boat deck. A piece of lamina could not be removed with any degree of accuracy. Therefore Point Haswell was chosen for the 1967 collection site because the kelp could easily be obtained in quantity under most weather conditions. On two occasions only in this locality, no samples were taken as high tides and stormy conditions made it impossible to reach the mid-sections of the stipes of the kelp. It should be noted however that combined analyses of both the short and long-term data indicated finally that the original "lucky-throw" method was the better collection method.

Temperature was the only external factor measured systematically. It is possible that the temperature of the water covering the surface laminae could be in the order of degrees warmer, or cooler, than that sampled in the bucket. The greatest degree of difference between the temperature of the water sample and that at the lamina surface is most likely to occur in calm weather, when the kelp floats virtually undisturbed. The well defined fluctuations that appear in the biweekly statistical analyses page 5

Plate I

A: Habitat of O. geniculata in Wellington Harbour, on the laminae of Macrocystis pyrifera.

B: Side view of Macrocystis lamina to show colonization by O. geniculata in winter.

C: An atypical erect stem taken in April showing a circlet of basal gonangia arising directly from the stolons.

D and E: Gonangia arising directly from the stolons. Sample collected in February.

F: Widely spaced summer stems showing decreasing order of vertical height.

page 6 (Figs. 4 and 5) possibly reflect sudden and short-lived temperature changes in the microclimate of surface laminae. However, calm conditions do not occur very often in Wellington Harbour. This is evidenced by meteorological records, and by the fact that of the 77 samples taken, only one sample had colonies of O. geniculata on one side of the lamina, and not on the other. No significant difference between colonization of the two surfaces of the lamina was found in any other sample. From this it seems reasonable to conclude that laminae have a fairly constant turnover by wind and water. Each side has an equal opportunity for colonization by settling planulae. It also indicates that with rare exceptions the water temperatures measured in the dip-bucket would closely approximate that at the lamina surfaces.

Fresh water influence, and pollution from the Ngauranga Freezing Works were found by Wear (1965) to have no marked effect upon the zoo-plankton in the collection area. Wear also found that the peak of the planktonic cycle occurs within the winter—May to August—period. Plankton is probably a major component of the food taken by Obelia. Moreover, it was anticipated that if these and other factors such as pH played a major role in influencing the growth rate of the hydroid, they would be indicated in the statistical analyses of the data.

Features of the life cycle of O. geniculata that bear on the interpretation of seasonal growth changes given here, but are masked in statistical analyses, are noted later in the text.

Terminology

Hammett & Hammett (1945) regard the word 'season', applying to summer, winter, spring and autumn, as a "waste-bucket word of loose usage", unless limited to the interval between a solar equinox and a solar solstice. For the present paper we have not defined the word with the exactitude of Hammett & Hammett. But we do define it specifically with regard to Wellington Harbour in terms of approximate temperature range, as follows: winter, 9°C-12.5°C; summer, 15.5°C-20°C; autumn and spring, 12.5°C-15.5°C.

There is no doubt that loose usage of the words winter, summer, etc. can be misleading. For example, peak production of Obelia medusae in Wellington Harbour is winter, namely July-August (= temperature 9°C-12.5°C, latitude, 41.2°S). In Perseverance Harbour, Campbell Is., it is summer, namely February-March (= temperature 9°C-10°C, latitude 52.3°S). Thus, if peak production of medusae is not also defined for Perseverance Harbour in terms of temperature and/or latitudinal range, it could imply a correlation between the season of peak production in this southern hemisphere locality, and the season of peak production recorded for northern hemisphere localities. Formation of gonangia reaches a peak through increasing warm water (summer) conditions in the northern hemisphere (Hammett & Hammett, 1945: Russell, 1953). In the southern hemisphere the peak is reached through decreasing temperatures. When seasonal terminology is disregarded, Perseverance Harbour is seen to be similar to other southern hemisphere localities in the New Zealand region where maximum production occurs in cold water temperatures about 9°C. That is, winter, in seasonal terms of the lower latitudinal range of Wellington Harbour waters.

Results

A. 1967 Autumn-Winter Season Text-Figs. 2-5

Statistical analysis—Temperature and growth correlation

The data obtained from the samples were analysed (Figs. 2-5) to determine if there were a relationship between temperature and three characteristic stem structures, namely, gonangia, terminal buds and axillary branches. Figure 2 expresses the relationship between temperature and gonangia. The Regression Curve equation page 7

Fig. 2 1967 autumn-winter seasons. Gonangia plotted against temperature using the statistical method of least squares.

Fig. 3 1967 autumn-winter seasons. Correlation between the average number of gonangia per stem, and temperature.

page 8 using the method of Heine (1966) is: Y = mX + c. The value of m = −0.45 and the value of c = 7.9. Therefore, Y = 7.9 − 0.45X. The correlation coefficient (r) for the gonangia is −0.94. Thus, the percentage of variability explained (r²) is 88%. Using a similar method, the percentage of variability explained for the terminal buds is 23% where Y = 121.28 − 5.56X, and for the branches 21.75% where Y = 83.62 − 4.73X. The same trend of increasing numbers of gongangia, etc., with decreasing temperature is seen using the simple graphic relationship for two variables (Figs. 3-5).

Remarks

Both series of statistical analyses demonstrate clearly that increasing numbers of gonangia, etc., are formed with decreasing temperature. Both series suggest (a) that in summer the production of gonangia, terminal buds and branches will fall to zero (Figs. 2-5), and (b) that temperature is a major factor in determining the trends of seasonal growth expressed by the erect stem (in terms of percentage of variability explained). Analysis by simple graphic relationship also demonstrates that great variation can occur in the percentage of stems possessing gonangia, terminal buds and branches within the weekly or monthly sampling period. Furthermore, this simple analysis suggests a critical temperature at which fluctuations in growth may take place. By critical temperature in this context, we mean that temperature above 13°C e.g., (Fig. 3) could either reduce or prohibit formation of gonangia or terminal buds. Temperatures above 11.5°C, e.g., (Fig 4) could either reduce or prohibit the formation of branches in the axil of the hydranth pedicel.

The annual range of morphology of the erect stem, described below, indicates however, that the relationship between stem growth and temperature is more complicated than is suggested by the results obtained from the autumn-winter seasons alone.

B. Annual Temperature and Growth Correlation
[Pls. I, II and III: Text-Figs. 6-9]

1961-1963 collection series

Plates I and II show the features of the erect stem for the four seasons. In winter (Pl. I, b: Pl. II, a) typical stems are close together, characteristically tall (15 to 20 mm), terminated by a hydranth bud, with 1 to 3 branches and up to 8 gonangia per stem. The stems also carry a fairly heavy epiphytic population of diatoms, and the chitin of the nodal annulations is medium to dark brown in colour. In summer (PL I, f: PL II, f) the stems are short (8.5 to 12 mm), wide apart, usually with a fully formed terminal hydranth, without branches, and with up to 2 gonangia per stem. They carry a heavy epiphytic population of diatoms. As could be anticipated, the spring and autumn stems (PL II, b and h) show features in between these two extremes. However, the spring stems frequently carry a greater epiphytic diatom population and more branches than the autumn stems (Pl III).

Statistical analysis

Figure 6 shows the pattern of increasing numbers of gonangia, terminal buds and branches with decreasing temperature, but does not indicate the well defined seasonal growth forms indicated for the erect stem in the visually selected examples of the stem described above. Figure 6 also demonstrates that gonangia and terminal buds may occur throughout the year. However, in late spring (November) or in the summer months (December, January, February), the percentage of stems with page 9

Plate II

A: Tall, branched winter stems.

B and D: Short spring stems, with epiphytic diatoms.

C and E: Medium height spring stems.

F: Summer stems with epiphytic diatoms.

G and H: Autumn stems of O. geniculata. The epizooite is a species of Epistylus which is commonly attached to the erect stem at this season.

page 10 branches is very low. The percentage may even reach zero. Moreover, in the two late spring, summer seasons analysed, there is a well defined rise in the production of gonangia. This production rise of gonangia is more or less paralleled by increased numbers of medusae in the plankton.

Some of the data of the 1961-1963 collection series is shown statistically expressed using the method of least squares. The calculated percentage of variability explained for temperature and gonangia (Fig. 7) is 46.04%, where y = 3.06 — 0.15X. That is, at temperature 12°C, forty-six percent of the stems would have 1.2 gonangia per stem. For gonangia and branches (Fig. 8) the percentage of variability explained is 46.04% and Y = 0.15 — 7.63X. Also, the percentage of variability explained for temperature and branches and for temperature and feeding polyps was calculated, but is not figured. The percentage for temperature and branches is 27.5% and y = 0.176 — 0.0148X. In marked contrast is the percentage for the number of feeding hydranths on the hydrocaulus. The percentage explained is only 0.4% where y = 5.53 — 0.024X.

Remarks

Visual assessment (Pls. I and II) and graphic analysis (Fig. 6) demonstrates a changing pattern for growth of the erect stem during the year. The highest percentage of terminal buds, branches and gonangia occur in the autumn to winter season. All these stem characters are found with decreasing frequency on the stem as water temperatures rise in spring and summer. Branches rarely occur on stems when water temperatures approach the maximum. Gonangia in contrast rise in numbers with the approach of maximum temperature. All in all, these trends suggest a strong correlation between temperature and the seasonal form and maturity attained by the stem. However, when this correlation is tested utilizing the method of least squares, the assumption that temperature is the controlling factor in seasonal habit is less obvious. The percentage of variability explained ranges from non-significant for the feeding polyps (0.4%) to varying degrees of significance, for the other structures.

For the branches and terminal buds it is 27%.

For the gonangia it is nearly twice as high, namely 46%.

In brief, the analysis of the long-term series indicates that changes in stem form occur throughout the year. These are related to the four seasons and thus indirectly to temperature. The degree of correlation between temperature varies for the stem characters assessed. It is lowest and non-significant for the feeding polyps. It is highest and quite significant for the gonangia. Correlation for the terminal buds and branches is significant and in between the two extremes.

Comparative Account of the Autumn-Winter and Annual Growth Trends

The autumn-winter seasons of 1967 and the annual analyses of the 1961-1963 seasons give a correlation between temperature and growth expression for the branches, and terminal buds as 21% and 27% respectively. While this is not a high degree of correlation it is acceptable statistically as significant for biological material. Nonetheless, it indicates that factors other than temperature are operative in determining the changing pattern of these stem structures, and that these factors operate at varying levels throughout the year.

In contrast is the correlation between temperature and gonangia for the autumn-winter seasons of 1967 and for the similar seasons of the 1961-1963 series. The percentage of variability explained for the autumn-winter seasons is 88% as against 46% for the whole year. In the autumn-winter seasons, with r² −0.94, the probability of there being no correlation is less than 1%. Temperature then accounts for page 11

Plate III

Photomicrograph of part of a late spring stem to show the epiphytic diatoms.

page 12 88% of all factors influencing the growth of the gonangia in autumn and winter. This is highly significant. In a biological system it is recognised that it is impossible to make a 100% correlation with any factor or factors. It is unlikely therefore that data on other external factors, e.g. pH, salinity, etc., would increase the coefficient by a significant amount. The margin of 0.006 left to be explained is very small, and could be an error in methodology.

The much lower correlation with temperature for the four season range, in comparison with the two seasons, is a reflection of two facts. Firstly, gonangia occur in small numbers throughout the year. Secondly, the numbers of gonangia produced rise sharply when the temperature approaches the annual maximum. Neither of these occurrences were predictable from the short-term results. The short-term data give a straight line by the method of least squares and suggest that gonangia numbers will drop to zero for temperatures above 17°G (Fig. 2). If the autumn-winter series had been the only series analysed, the result, as already noted, would have been strongly biased in favour of an exogenous factor being the controlling agent of the stem form attained by O. geniculata. Sampling over a whole year clearly shows that a sigmoid curve is a better representation for the seasonal trends in gonangia formation. Nonetheless, the results for the terminal buds, the branches and the gonangia demonstrate that temperature is significant in the overall growth pattern of the hydroid. It is particularly significant in the formation of the gonangia.

The results obtained from study of gonangial growth are recognised here as a more reliable guide than other stem features to the factors governing growth in O. geniculata. The reason is as follows. Where a series of samples are being utilized over a period of time, rather than observations on one colony, or one erect stem, it is best to have colonies at approximately the same stage of maturity. The presence of gonangia usually indicates that a colony is well established. Therefore their presence will also give a better indication than the number of hydranths per stem or the presence of terminal buds, etc. that a colony is mature. Moreover, hydranth development from bud initiation to senility is completed within 72 hours (Hammett, 1943, p. 350). Gonangial development including the initiation and development of medusae is not reached within 72 hours (Hammett, 1943) and may take several weeks. As the blastostyle in the gonangium reaches senility, it is shed and not replaced. In contrast, a hydranth may be produced from a site where a polyp had previously emerged, matured and regressed, and recurrent growth is integrated into the growth of the colony as a whole (Hammett, 1943). This growth cycle, characteristic of the feeding hydranth, undoubtedly contributes to the very low non-significant correlation between temperature and the number of feeding polyps present on the stem. Regeneration of hydranths in already present hydrothecae could mask almost completely the true relationship of temperature to this stem growth feature in long-term field samples. The gonangia therefore with their longer time of development and lack of regenerative powers are better stem structures for statistical analysis in the present instance. Coupled with the longer development time of the gonangia is the possibility that they do not respond to the same extent to minor day to day temperature fluctuations as do the hydranths and terminal buds. This is substantiated in the statistical analysis of the 1967 autumn-winter data where the time interval beween collections was the shortest for the present study. The fluctuations in successive biweekly samples (Figs. 4 and 5) are more clearly defined for terminal buds and branches than for gonangia (Fig. 3).

There are other featuŕes of the long-term series that bear on our interpretation of seasonal growth that have not yet been discussed. For example, gonangia and medusae occur throughout the year. It is probable therefore, that planulae settle and new colonies arise throughout the year. Thus, the monthly mean for any expression page 13

Fig. 4 1967 autumn-winter seasons. Correlation between the average number of branches per stem, and temperature.

Fig. 5 1967 autumn-winter seasons. Correlation between the percentage of terminal buds present and temperature.

page 14 of growth for the erect stem does not necessary represent the true average growth pattern of stems growing for one month at the mean monthly temperature. Individual samples for the month might represent a mixture of tall, mature stems; short immature stems, and stems of medium height and maturity depending on the time the stems have been growing from planula settlement. Alternately, the monthly average may represent individual samples in which one of the three stem types is dominant. The average percentage of gonangia shown in Figure 6 for November 1962 illustrates this situation. In one sample, 96% of the stems had gonangia present, were tall stems, with branches and with nodal rings of dark brown chitin. The stems and hydrothecae

Fig. 6 1961-1963 seasonal changes in the percentage of gonangia, medusae, branches and terminal buds relative to changes in temperature.

were thick with diatoms. All the characters of these stems were indicative of approaching senescence following a long growth period in cold water. The second sample had small stems, 100% of which lacked gonangia, and overall the stems were indicative of a short period of growth. The third sample, was a mixture of the above growth forms, namely the taller stems in the main had gonangia, while the majority of short stems lacked gonangia.

It has already been noted that there are two peaks in the annual production of gonangia, one in the winter and one in summer. But the stems that carry gonangia in summer are very different from those bearing gonangia in winter. This is another example where the bare statistic does not fully represent the seasonal morphology of the animal. Direct observation of the samples for February shows that most stems at this time of year are very short, seldom have more than 4 hydranths, and that the distal end is terminated not by a bud, but by a fully formed hydranth. There may also be one or two gonangia present. However, other patterns of stem growth along the stolon have been observed. Plate I, f, shows a decreasing order of vertical height for stems along the stolon, and the "erect stem" may be represented only by a page 15gonangium arising directly from the stolon (Pl. I, d and e). The latter "stems" however were not included in the statistical analyses of the data.

A variation of this pattern is seen in Pl. I, c., where the erect stem growth expression is shown by a single stem surrounded by a circlet of gonangia arising directly from short, radially arranged stolons. This example is of interest as it comes from an autumn collection. The pattern of numerous gonangia on the stolons in this sample suggests a sharp upward fluctuation in temperature at a critical level, combined with calm weather, resulting in near lethal temperature conditions at the lamina surface. That is, summer conditions in autumn.

Figure 6 shows what we regard as another example of the effect of unseasonable climate. In the March-April period for 1962, temperature fell from a mean maximum of 19.5°C in February, to a mean of 12.8°C in April. The lowest temperature recorded for April was 10.4°C. The latter temperature approaches the annual minimum for Wellington Harbour. A temperature range of approximately 10°C to 13°C usually occurs from June to August not March to April. It is probable therefore, that this rapid, unseasonable temperature drop is responsible for the very high percentage of stems with terminal buds in March and April instead of later in the year as is suggested by analysis of other data obtained during the experiment. The more gradual fall in temperature from the February maximum in the 1963 March-April period shows an appreciably lower percentage of stems with terminal buds.

In brief, the above paragraphs indicate firstly, that assessment of part only of the seasonal data gives a biased result in favour of temperature as the prime factor in the seasonal growth form attained by the erect stem. Secondly, that some facets of stem growth are also obscured in statistical analyses of the annual data. But a combined assessment strongly suggests the same seasonal trend in changing growth form for the stem as that noted in direct visual observation of the two collection series.

Fig. 7 1961-1963 sample series. Gonangia plotted against temperature using the statistical method of least squares.

Fig. 8 1961-1963 sample series. The percentage of medusae per total volume of catch, plotted against temperature using the statistical method of least squares.

Fig. 9 1961-1963 sample series. Average number of gonothecae plotted against the average number of branches per stem using the statistical method of least squares.

Seasonal Occurrence of Obelia Medusae in Wellington Harbour

So far, description has been of the colonial phase of the life cycle. Assessment of the seasonal occurrence of the medusa in the plankton has also a place in the final interpretation of stem growth. Obelia medusae are found all the year round in Wellington Harbour. Continuous occurrence is also recorded for the British Isles (Russell, 1953). Around British coasts, Obelia medusae are most abundant from spring to late autumn. In Wellington Harbour, they are most abundant from late autumn to winter. Numbers of medusae in the latter harbour do however also rise on occasion during summer (Fig. 6). But there is good evidence that this is due to aberrant growth causing the formation of gonangia on the hydrorhiza in place of the normal hydrocaulus. Thus, it is not a comparable rise in numbers to that found each winter. This aberrant summer rise is reflected in the annual low percentage of variability explained for the medusae per volume of catch when plotted against temperature. The percentage explained is 34.4%, where Y = 100.26 − 4.59X. (Fig. 8.)

Further evidence that there is a reversal in the southern hemisphere of the seasonal peak recorded for the northern hemisphere comes from records of Obelia from Perseverance Harbour, Campbell Island (latitude 52.3°S). Medusae are first recorded in January, when the sea-surface temperatures average approximately 10.0°C. A peak is reached in March (temperature average, 9.0°C). The medusae number is then between five and six thousand per 15 minute tow. By May, temperatures have dropped to about 7.0°C and the numbers of medusae per tow decrease, until by June no medusae were recorded (P. Roberts, Zoology Dept., V.U.W.—personal communication). The lowest temperature in this subantarctic area is approximately 5.0°C recorded in July-August.

The temperatures at which medusae have been recorded from Perseverance Harbour are comparable with those from Wellington Harbour in winter, where the lowest temperature is about 9.0°C. Wear (1965) found the peak in planktonic medusae in the latter area to occur in July-August. From both plankton studies it appears then that the largest numbers of medusae are obtained in the New Zealand region when the sea surface temperatures are around 9.0°C. It is also probable that there is a lower lethal temperature. This may occur in areas where the sea surface temperatures fall markedly below 9.0°C for any length of time.

Interpretation of Results

Our interpretation of the overall growth sequence for Obelia geniculata from the settling of the planula, is nutrition, reproduction and regression. We regard this sequence as genetically determined, and not alterable by exogenous influences. But the rate at which the sequences proceed is materially affected by the season at which the planula settles and thus indirectly by temperature. Consequent on this conclusion, is the fact that the developmental growth form finally attained by the erect stem is variable from season to season and within the season, according to the length of time the stem has been growing within a particular range of temperature. This hypothesis is shown diagramatically in figures 10 and 11.

Support for the idea that exogenous factors such as temperature affect the growth processes of hydroid colonies comes from the work of Berrill (1948). Berrill found that growth in colonies of Obelia and other hydroids fluctuated greatly with temperature. They disappeared and reappeared respectively as the temperature rose and fell significantly above and below 20°C. One of the three species of Obelia studied by Berrill was O. geniculata. It is probable therefore in latitudes where the mean annual sea temperature is approximately 20°C that O. geniculata will be seasonal in its growth. Temperature is thus not a limiting factor in Obelia growth in Wellington page 18

Fig. 10 Diagrammatic presentation of the hypothesis that seasonal changes in the exogenous factor temperature, is related to the metabolic rate and growth of the erect stem.

Harbour, except perhaps in February, when the microclimate of the surface laminae may exceed 20°C in days of calm hot weather. Also, growth is unlikely to be seasonal in latitudes in the New Zealand area higher than 40°S.

It is also concluded from the present study, that increasing temperature produces an increase in the metabolic rate of the colony, and that the endogenous sequences, nutrition, reproduction and regression, follow one another in rapid succession. Of significance for this conclusion is the decreasing order of stem height illustrated in Plate I, f., because as Manton (1940, p. 248) notes, budding in colonial hydroids takes place when growth is complete. This growth pattern of decreasing height along the stolon, also suggests that the rate of metabolism may be sufficiently rapid when temperatures approach the upper lethal limit for the two growth sequences, nutrition and reproduction, to be virtually telescoped one into the other. The "erect stem" then consists of a gonangium arising directly from the stolon (Pl. I, d and e). This type of "stem", is most likely to occur in Wellington Harbour if the planula settles in a period of high (18°-20°C) temperature. Nutrition for the developing gonangium would be provided by the hydranth-bearing stems already present on the stolon.

Fig. 11 Diagrammatic presentation of the hypothesis that the seasonal stem form attained, is related to the season at which the planula settles and the time available for growth within the seasonal temperature range.

If the planula settles earlier in the summer season, it will have a longer growing time in cooler water, and the stem will be able to produce feeding polyps in varying number, the number attained being dependent on the temperature when the planula settled and the subsequent growing time available before lethal temperature is reached.

In winter the metabolic rate slows with a decrease in temperature. The time period for the sequence nutrition, reproduction, regression, is lengthened. Before gonangia are produced the stem attains a vertical height approximately twice that attained by a stem in summer. If the planulae settle in a month of minimum temperature range the stems will give expression to growth, not only in vertical height, but also in the production of branches. Further evidence for this conclusion is given by the stem form of colonies collected in February from Kerguelen Island (Ralph, 1956). Mean annual sea temperature for Kerguelen Island is approximately 2°C below minimum winter temperature of 9°G in Wellington Harbour. No gonangia are present on these sub-antarctic colonies, but 80% of the stems have long branches. The erect stems without branches are very short, with terminal growing buds suggesting that page 20 they were young stems. The length of some branches approximated that of the main stem, and we regard this growth pattern as an indication that absolute vertical height is genetically governed. Moreover, it is possible that this vertical height may be attained in advance of the endogenous control mechanism initiating reproduction when long slow growth is possible. The nutritive growth phase is expressed then in the form of an axial bud which forms a branch and not a gonangium. The correlation (46% of variability explained) between gonangium and branches further supports this conclusion. Slow growth in cold waters has been postulated for other hydroid coelenterates, for example Myriothela penola from the Argentine Islands area of Graham Land (Manton, 1940).

Other marine invertebrates in New Zealand waters are known in which there is a significant size difference in breeding animals in warm and cool water areas of the distribution range. For example, the females of the spiny lobster Jasus edwardsii (Hutton, 1875), come into berry in warm east coast, North Island areas at a smaller size than the females in the south-eastern and southern regions of the South Island. (Dr. R. B. Pike, Zoology Dept., V.U.W.—personal communication).

If it is assumed that the above hypothesis concerning fast growth in summer and slow growth in winter is correct, then it follows that there will be a critical temperature in autumn and spring at which temperatures below will slow down the overall stem metabolism. Such temperatures are indicated in the statistical analysis of the autumn winter data for 1967, i.e. approximately 13°C for terminal buds and gonangia and 11.5°C for branches.

We do not know the length of the stem life cycle of O. geniculata, but have concluded from the present study that the length is temperature dependent in large measure. Therefore it would be "short" in summer and "long" in winter. It is also probable that it is longer still in latitudes significantly higher than Wellington Harbour. For example, Kerguelen Island. It could be indefinite in this latitude. The feeding hydranths may be able to survive much longer before regression sets in. They may not have a cycle of regression followed by regeneration at all. Once fully formed and functional they may remain as such until the whole erect stem regresses. It is not inconceivable that regeneration of the blastostyle also occurs in such latitudes. Furthermore, it is possible that regression of the feeding hydranth in winter temperatures in Wellington Harbour is not followed by regeneration to any great extent, and that the reverse is the case in summer temperatures. For example in summer, the number of feeding polyps during the life of the stem could be the same number as in winter, if there was a very rapid succession of feeding polyps regenerating in already existing hydrothecae.

As is so often the case with studies of the present nature, while some questions may be answered, many more arise and cannot be answered from the data obtained in the originating experiment. Future laboratory studies could provide some answers to the questions posed by the present field work.