Morphometric studies of Sargassum spp. (Sargassaceae, Pbaeophyta) from Reunion Rocks, KwaZulu-Natal, South Africa. I. Receptacles

Specimens of a mixed population of Sargassum spp. were collected from Reunion Rocks , KwaZulu-Nalal , South Africa over a period of one year. Three general shapes of receptacle were recorded for Sargassum elegans, S. incisifofium and an undescribed species of Sargassum: i) terete , ii) three-cornered and iii) twisted receptacles respectively. The number of conceptacles on the receptacles was enumerated and rela ted to the dry weight of the receptacles. It was found that conceptacJe number per receptacle correlated well to the weight of the receptacle and that each receptacle type had a different number of conceptacles per unit weight. Re lationships were found to vary from one season to the next, however, it was found that the three-cornered receptacle type developed a greater number of conceptacles per un it weight than the terete receptacle type. Reproductive:somatic biomass ratios indicated that the three-cornered receptacle type had a lower reproductive expenditure than the terete receptacle type in terms of biomass. Ecological and possible taxonomic implications are discussed.


Introduction
The brown algal genus Sargasstlln has a tropical to sub-tropical distribution world-wide. The genus is cred ited with over 500 recorded species (Yoshida 1989), but the taxonomy of the genus is considered by several authors to be in need of revision (Tseng el al. 1985;Yosh id a 1988Yosh id a ,1989Abbott 1992). There have been many papers which have in vest igated the taxonomy of Sm-go.\"slim spp. (e.g. Yoshida 1989;Kilar el al. 1992a,b;Trona 1992;Tseng & Lu 1992a.b;Ajisaka el 01. 1994;Noro el al. 1994;Silva el al. 1996) and several papers have investigated aspects of the ecology of members of the genus (e. g. Ang 1987;Ang & de Wreede 1990;Trona & Tolentino 1993;Kendrick & Walker 1~94 and Largo el al. 1994) . Some papers have attempted to place a measure on the reproductive output of the species under investigation (Umezaki 1984) and others have high lighted the reproductive development of me mbers of the genus (Okuda & Satoh 1989;Critchley el al. 1991). To our knowledge, few papers have related quantitative data on reproductive structures across different species/forms of SmgasslIm.
Thi s paper for ms part of a broader ecological study of the Sa/"-gasslllJl spp. population at Reunion Rocks, KwaZulu-Natal (KZN), South Africa. Sargassllm is a common genus of seaweed occurring on the coast of KZN and is a dominant alga of upper intertidal rock pools. The SargasslIm population at Reunion Rocks (approximately 15 km south of Durban; 29°59'OO"S, 30°58'40"£) consists of several forms, the taxonomy of which remain un certain (Birkett 1984). Several taxonomi c publications have included Sout h Afr ican species of Sargassllm (Simons 1976;Seagrief 1984), and nine species have been recorded from the Sou th African coast (Silva et af. 1996), yet the variation present at Reunion Rocks is not adequately described by these works. Three for ms of Sargassllm, each with a different shape of receptacle, are disc ussed in this paper. One is Sargassum elegans Suhr. one has been commonly referred to as S. helerophyllul1l C.
The reproductive structures of Sargassllm are referred to as receptacl es. In the forms found at Reunion Rocks, these structures occur in the axils of laminae of either secondary or tertiary branches of the main axis (nomenclature after Jensen 1974 and Yoshida 1989). Conceptacles (structures within the receptacle containing gametangia) occur in the receptacles and are visible on the receptacle surface as pores.
The number of conceptac les that eall occur on a single ret:eptacle is considerable and an accurate coullt of these cOllceptades while rotating a receptacle under a mi croscope is both difficult and time cons uming. This study presents an ana lys is of all ometric relati onships which allow estimation of conceptacle number from receptacle weight, while, at the same time presenting useful quantitative information of reproductive features of the genus. often considered to be conservative characters, across di ffere nt species within the genus Sargasslflll. The potential application of a quantifiable taxonomic character and some ecological implications of the results are discussed.

Methods and materials
Thalli of SargasslIlIl eiegal1s, 5,-II/Cis/folium and an , as yd. undescribed Sargassum spcdc-s w~re COIlCClc-d from Reunion Rocks, KwaZulu-Natal , Smllh Africa (29°59'OQ"S. 30 0 58'40''E) over " one year period (July. Dctuber 1995, J8.lluary. May 1996). E8.ch s peci~s develops reccptacles of a differcnt fo rm. a consistent feature belween the sredes. Stratified random sampling using sCV(:1\ 0.16 m 2 quadrats was used . From the material collected. a sub-sample ora minimum of five thall i (with recept acles) of each species present was chosen at random. From these thalli. all receptacles present were removed and preserved in 4% form alin-sea\vater. A sub-sample of receptacle clusters was taken from the thall i. A receptacle duster being delined as all those reccptad: hranch t::s (a ll three types of receptaclt! display branching) which arb!.! ti·olll onc axis ill th!.! axil of a lamina. The numba of conceptacl!.!s thaI could be seen on a n.:ee ptacle when look ing down the mi croscope was counted fo r matt:rial collected du ring all sample ti mes. i.e. one side of the thrc-t:-cornered receptacles (receptacle placed on (lll~ apex o f the ·triangle '); those conceptaclcs that wen: visihle on the lereh: typL' without any movement of the receptacle, and those of the twisted receptach: were counted by following one of the three sides around a twist from the base to the apex. In additio n to th!.!sL' data tbe lo tal number of COI1ceptacles per receptacle cluster was also counted for Ju ly 1995 so as to determinc the relationship hetween ' surface' and total conccptacles. Th e receptacles wt.!rc examined using a Wild skrl.!O d issect in g Ililcroscll pe. The shape. number and dry weig.ht of n:ccptacks were rewrded along with the weights of the associated thalli . The receptar.:ies and thalli were dried at 45°C lllltil constant 357 weight. Regression anal~ ses \\ ere performed HI detenn inc rdat ioll w ships between the variolls parameters.
These names refer to the general shape of the receptacles. This paper does not intend to describe the nature of the receptacles in detail , but rather indicate the different forms that do ex ist and which will be discussed. The terete form lacks ornamentation and is approximately circular in cross-section. The twi sted type has three 'sides' with a spiral nature. T hi s receptacle type has more ornamentation (in the form of po inted protuberances) than the three-cornered type, which. when cut in transverse section presents the forrn of a triangle. T he three-cornered receptacle type lacks ornamentation, but for the corrugated surface whi ch occurs as a result of the occurrence of cOll ceptacl es. A II three receptacle types were fou nd to contai n oogonia throughollt the year.
'Surface' conceptacle num ber was plotted against total conceptacle number fo r each rece ptacle type collected during Jul y 1995. Correlation coelficients (r) 0 1"0.997_ 0.997 and 0.984 were obtained for the terete, three-cornered and twisted receptacle types respectively ( Figure 2).
For subsequent ana lyses 'surface ' conceptacles were co unted and from this the total conccptacle number for each conceptacle type was estirnated using the formulae presented in Figure 2. It was believed that thi s approach. while in addition to being more efficient, would al so yie ld acc urate res ults as counting recepta w cles around an entire receptacle can result in errors associated with recounting of conceptacles when rotating the receptacle to view all sides. This would be more noticeable when counting the terete and twisted receptacle ty pes.
In general, the three-cornered reccptllcle type devel oped a greater number of conceptacles per receptacle than the other two receptacle types (Figure 2). The maximum number of concepta-cJes recorded from a three-corne red rt;ceptacle cluster exceeded that of the terete receptacle cl usters for the four seasons studied.
The g reatest difference was recorded in July 1995, where a receptacle c luster of the three-cornered type had 1559 conceptacles, while the m ax imu m number of cOllceptacles recorded from a terete receptacle duster was 632. The sig nificance of this became clearer when the number of conceptacles was related to receptacle weight (Figures 3-6) . Regression analyses revealed a strong'relat ionsh ip between the number of COIlceptacles on a receptacle and receptacle we ight. What is apparent from Figures 3-6 is that each of the rcceptnclc types had a different relati onship of conceptacie number to receptacle weight. As a rule, the twisted receptacle type hnd a lower number of cOllccptacles per weig ht of receptacle than the ot her two types. The three-cornered receptacle type had the highest number of conceptacles per wei ght of receptacle. The terete receptacle type is intermediate between the two . This is indicated by the s lopes of the lines, III ofy = IJI.\" + C, presented in Fig ures 3-6.
Although the regression analyses revealed a relationship of conceptacle number to receptacle weight for each of the seaso ns, the relationship for each receplacle type and the relation ship of one receptacle type to another did not re ma in consta nt for each season as indicated by the different regressio n equatio ns for each seaso n (Figures 3-6). A general relationship was identified, i.e. per unit of receptacl e weight the three-cornered receptacle type S. Afr. 1. Bot. 1997,63 (6)   developed a greater number of conceptacles than the terete receptacle type; 111 for the three-cornered receptacle types was cons istently above III for the terete receptacle type. It is difficult to make statements concerning the twisted receptacle type as data fo r this receptacle type is limited. Although thall i with these receptacles were fou nd, their nu mbers throughout the sample 1600 1400 ... .. 1200 (]) .0 tha ll i w hic h dt:vcl o ped three-cornered receptacles prod uced receptacles wh iell cons ti tuted a sma ll er proportion of th e somati c biomass than the other receptacle types (Figure 7) .

Discussion
Chapm an (1995) warns th at th e estim ati on of fertility from reprod uctive or vegetat ive bi omass is prone to e rror.  • s. Atr. J. Bol. 1997,63 (6) • 3-cornered receptacles  which dev elo p tht.:: twisted receptacle type (Sargass lI171 sp.) put more energy into developing receptacles with fewer concepta-c1es. T hnlli whi ch develop the three-cornered receptacle type (s. im:isi{oliltl1l ) invest th e least energy in receptacle production than the other receptacle types in terms of assoc iated somatic biomass. However, the receptac les have a greater number of conct!ptaclcs, the units of gametangia production, per unit of rcceptacle weight. Observations indicated that thalli with the twi sted receptacle type were the least abundant at Reunion Rocks, whi le S. il1cisifo/iut11 and S. e/egalls see med to be more ba lanced in terms of abun dance (unpublished observations). The two species can grow alongside each other, however, thalli with terete rece ptacles (.")' argasslIlI1 elegal1s) tend to occur in more wave exposed habitats than th alli with three-cornered receptac les (S iJJcis{fo/iwII ). Th e differences in conceptacle number per unit receptacle weight have a poss ible ecological significan ce and Illay represent different reproductive strategies. It is possible that more conceptacies are produced per unit receptacle weight to ensure a greater chance of fertilization under con ditions of greater wave actioll. De Pau la and De Ol ivei ra ( 1982) found that the receptac les of S cym oSUnT were more branched in areas of greater surf exposure and suggested that this in conibination with short leaves may increase the possib ility of ferti li zat ion un der surf conditions. Birkett (1984) revealed that of the forms of Sorgos.HlIIl at Re un ion Roc ks, published description s of Sou th Afr ican species of SargasslIl11 could not be co nfidently applied to all of thos e present. Observations of herbari um spec imens and study of desc riptions have not led to the pos itive ide ntification of all the forms present at Reunion Roc ks and it has been suggested that several of the form~ present may be undescribed species (Birkett 198 4) and some may be morphological variants of a single species. Three species of Sargas.mlll are d iscussed in this pape r, .\'. e/ef{ans, 5; il1cis(Fbliltlll and an, as yet, undescri bed species.
Although thalli with different receptacle types appear to dominate in different micro-hab itats at Reunion Rocks, poss ibl y related to wave exposure, they can be found growing alongside each other se parated by distances of only a few centimetres. The di ffere nces identified, in terms of several relati onshi ps di scussed. between S. inc:isifh/iul11 and S. elegal1s appear to be relatively temporally stab le. The differences identified m ay be an ind icati on that the thalli are separate taxonomi c entities and these quantifiab le features of reprod ucti on warrant furt her investigat ion as taxon omic characters.