M. T. Brown AE M. A. Nyman AE J. A. Keogh N. K. M. Chin Seasonal growth of the giant kelp Macrocystis pyrifera in New Zealand Received: 16 May 1997 /Accepted: 28 May 1997 Abstract The seasonal growth pattern of Macrocystis pyrifera (L.) C. Agardh in New Zealand was determined from measurements of blade-elongation rates between March 1986 and December 1987. Growth rates re- mained relatively constant throughout most of the year, but declined significantly during the summer months. Seawater nitrate levels had a marked seasonal cycle, with concentrations of 0.3 lM detected in summer. The timing of the decline in internal nitrogen concentrations varied for di��erent-aged blades, but occurred 1 mo after the decrease in seawater nitrate concentrations. Su��ciently high irradiance levels and seawater nutrient concentrations support active growth for most of the year, but growth is nutrient-limited during summer. No carbon accumulation during summer was observed. Between March and August 1988, growth estimates were also derived from node-initiation rates and stipe-elon- gation rates to permit comparisons with previous studies from the northern and southern hemispheres. Direct comparisons of the three growth parameters determined for non-terminated canopy and sub-surface fronds were used to assess possible di��erential resource allocation in the two frond classes. Introduction The giant kelp Macrocystis pyrifera (L.) C. Agardh has a bipolar distribution, with circum-antarctic and north- east Pacific components it occurs in cold temperate latitudes, but also extends into lower latitudes on the Pacific coast of South and North America, in areas of major coastal upwelling (Chin et al. 1991). Although M. pyrifera is of greater ecological significance and has a wider latitudinal distribution in the southern hemisphere (Kain 1989), most of our existing knowledge on the bi- ology of the species comes from studies undertaken on the kelp beds o�� California, USA, where it is commer- cially exploited for alginate production (North 1986). In New Zealand, M. pyrifera grows in southern and central open coastal waters, mainly along the east coast. Its northern limit is between Castle Point and Cape Turn- again on the east coast of North Island and Kapiti Is- land on the west coast of North Island (Hay 1990). Hay considered its distribution on the coast of mainland New Zealand to be perfectly correlated with areas influenced by the Southland Current (Heath 1972, 1975). As well as being encountered on the outer coast, M. pyrifera is also to be found in sheltered harbours and enbayments, where it can form extensive fringing beds in shallow waters. In very sheltered coves of Paterson Inlet, Stewart Island, fragments of M. pyrifera grow vegetatively over a muddy substrate (Moore 1942 Gerard and Kirkman 1984). Only limited information is available on the in- digenous populations of the giant kelp in New Zealand, including its biogeography (Rapson et al. 1942 Hay 1990 Chin et al. 1991), morphology and development (Kain 1982 Nyman et al. 1990, 1993), and alginate chemistry (McKee et al. 1992). Studies on seasonal growth patterns are lacking. Seasonal changes in growth and productivity of Macrocystis spp. have been studied extensively in Cali- fornia (e.g. Wheeler and North 1981 Dean and Jacob- son 1984 Zimmerman and Kremer 1986) and British Columbia (e.g. Smith et al. 1983 Wheeler and Druehl Marine Biology (1997) 129: 417���424 �� Springer-Verlag 1997 Communicated by J.P. Thorpe, Port Erin M.T. Brown (&)1 AE J.A. Keogh AE N.K.M. Chin Department of Botany, University of Otago, P.O. Box 56, Dunedin, New Zealand M.A. Nyman Department of Mathematics and Computer Science, Alma College, Alma, Michigan 48801-1599, USA Present address: 1Marine Biology and Ecotoxicology Group, Department of Biological Sciences and Plymouth Environmental Research Centre, University of Plymouth, Plymouth PL4 8AA, Devon, England

1986), whereas only two such studies have been under- taken in the southern hemisphere, the Kerguelen Islands (Asensi et al. 1981) and Falkland Islands (van Tus- senbroek 1989). From the results of all these studies, it is evident that there is no generalised growth response for this genus, but rather that variations in local environ- mental conditions a��ect the pattern of growth. Macrocystis pyrifera is a morphologically complex seaweed, with new fronds being produced from basal meristems which di��erentiate into stipe and blades from an apical meristem. Increase in stipe length is due to growth in the immature portion towards the distal end of the frond, and lamina elongation occurs by the ac- tivity of a meristematic zone located at the lamina/ pneumatocyst transition-zone. Thus, estimates of growth can be obtained from several di��erent parame- ters. Commonly, frond elongation rate, the increase in length of stipes with time, and node initiation rate, the increase in number of nodes on a frond with time, are used (e.g. Zimmerman and Kremer 1986 van Tussen- broek 1989 Gonzalez-Fragoso et al. 1991) but frond- initiation rate, increase in number of fronds with time (Zimmerman and Kremer 1986), and changes in weight of fronds over time (Wheeler and Druehl 1986) have also provided estimates of growth. In this study, growth rates, estimated from elongation of blades, were measured over a 2 yr period and related to observed fluctuations in environmental variables. The chemical composition of the blades was also determined, and this too was related to temporal changes in growth rates and environmental conditions. In the third year of the study, growth estimates obtained from measure- ments of blade and stipe elongation and node-initiation rates were compared, and di��erences between canopy and subcanopy fronds were assessed. Kain (1989) suggested that Macrocystis pyrifera can be classified as a ������responder������ rather than an ������antici- pator������, i.e. growing and reproducing when environ- mental conditions are favourable. At high latitudes (50��) growth is highly seasonal and correlated with available light levels, whereas at lower latitudes (30��) the species is aseasonal. Our results, obtained from an intermediate latitude (���45��S), will be discussed in the context of this hypothesis. Materials and methods Site description The study site is located close to the Portobello Marine Laboratory of the University of Otago at Aquarium Point, Otago Harbour, on the east coast of South Island, New Zealand (170��50��E 45��40��S). Otago Harbour is long and narrow, lying approximately NE���SW, with the city of Dunedin at the inner end and a constricted outlet at the other. Tides are semi-diurnal with a range of 1.7 m at mean neaps and 2 m at mean springs. The Macrocystis pyrifera bed is situated in shallow water, with holdfasts ���3 m below mean low- water springs, adjacent to a constricted tidal channel with currents up to 5 m s)1 during tidal movements. Blade elongation Each month, between March 1986 and December 1987, five healthy non-terminated canopy fronds from three di��erent plants were se- lected, and three blades on each frond, representing di��erent stages of maturity, were tagged elongation rates were measured over a 30 d time interval. Both plants and fronds were selected at random. Blade types were characterised by appearance and position on fronds following the definitions outlined by Coon (1981) and Lob- ban (1987): immature blades = first free blade from the apical scimitar of each frond mature blades = 10 blades below the im- mature blade, more pigmented and corrugated than immature blades old blade = 10 blades below the mature blade, with extens- ive bryozoan growth on blade surface and much distal tissue decay. Elongation rates were determined using the punched-hole technique of Parke (1948) a series of 5 mm holes were punched in blades, with the first hole 2 cm from the pneumatocyst/lamina transition-region and at 10 cm intervals thereafter. The distance from the last hole to the tip of the blade and total blade length were also recorded. After 30 d, the intervals between the series of holes were remeasured and the growth and/or the amount of distal ero- sion was estimated. Relative growth rate, RGR (Evans 1972), be- tween holes was calculated as: RGR = ln ��D2D1 1�� ��T2 T1��, where D = distance between the same two holes and T = time. Frond elongation Estimates of frond-elongation rates (change in stipe length per unit time) and node-initiation rates (change in number of nodes per unit time) were determined for all non-terminated fronds of three ran- domly selected plants at monthly intervals between March and August 1988. Each frond was tagged at the base and also just below the apical scimitar. In addition, lengths of all free blades on selected fronds were recorded and changes in the total length of blade material were calculated. As new fronds appeared and blades de- tached from the apical scimitar, they were tagged and lengths re- corded. In addition to assessing temporal changes, spatial di��erences, i.e. position in the water column, were determined by comparing growth data of canopy (��� 2.25 m in length) and sub- merged fronds. Chemical composition Blades used for in situ growth estimates and immature blades re- cently detached from apical scimitars were harvested at the end of each 4 wk growth period during the first 14 mo of the study and analysed for carbon and nitrogen content. Tissue was freeze-dried and finely ground in a Wiley mill. Concentrations of carbon and nitrogen were determined using a Perkin-Elmer Model 29-B ele- mental analyzer. Samples, run twice, indicated an instrument variation of 0.01 to 0.05%. Environmental parameters During monthly tagging and retrieving of blades, which usually took place during low tide, three replicate water samples were collected for determination of nitrate, ammonium and phosphate concen- trations and analysed using a nutrient autoanalyser (CHEMLAB Instruments Ltd.). Surface quantum irradiance was calculated from mean monthly irradiance data provided by the Dunedin Meteoro- logical Service, and mean monthly surface seawater temperatures were calculated from daily (09:00 hrs) readings obtained from the Portobello Marine Laboratory temperature database. Statistical analyses Data were analysed using JMPTM (SAS Institute, Cary, North Carolina). Where appropriate, non-parametric ANOVA and 418