182 S.AfrJ.Bot., 1992, 58(3): 182-187 Ontogenetic and demographic studies of grandiflora subspecies fimbriata ()

C. Ruiters,* B. McKenzie and L.M. Raitt Botany Department, Faculty of Science, University of the Western Cape, Private Bag X17, Bellville, 7535 Republic of South Africa

Received 18 September 1991; accepted 17 March 1992

The seasonal ontogeny and phenology of spp. fimbriata were examined along with various population parameters for this synanthous geophyte which possesses an annual storage organ. It was found that the majority of in the sample were less than seven years old with a maximum age determined at eighteen years. Age was successfully determined by examining subterranean markers (tunics) and it was found that the plants do not generally flower before they are four years old and have a 'critical minimum biomass' of 0.37 9 dry matter. Corm size is related to depth of burial until reproductive maturity is reached, with over 80% of the found in the organic profile which extends to 3 cm below the soil surface. The seasonal allocation of dry matter to the various parts indicated the efficient mechanisms adopted by this species to cope with the mediterranean climatic conditions and the low soil nutrient levels.

Die seisoenale ontogenie en fenologie van Sparaxis grandiflora spp. fimbriata, 'n sinantiese geofiet wat 'n eenjarige stoororgaan besit, is tesame met verskeie populasieparameters ondersoek. Daar is gevind dat die meerderheid van die plante in die monster jonger as sewe jaar oud was met 'n maksimum ouderdom van agtien jaar. Die ouderdom is suksesvol vasgestel deur ondergrondse merkers (tunikums) te ondersoek, en daar is gevind dat plante nie voor die ouderdom van vier jaar blom nie en dat 'n 'kritiese minimum knolbiomassa' van 0.37 9 daarvoor nodig is. Verder is daar ook gevind dat 'n verwantskap bestaan tussen die diepte waarby die knol voorkom en die knolgrootte, aangesien knolgrootte met diepte toeneem totdat reproduktiewe volwassenheid bereik is. Meer as 80% van die versamelde knolle het in die organiese profiel wat tot 3 cm onder die grondoppervlak strek, voorgekom. Die seisoenale verspreiding van droa massa na die onderskeie plantdele dui die baie goeie meganismes aan wat hierdie spesies ontwikkel het om tred te hou met die mediterreense klimaatstoestande en die lae minerale vlakke in die grond.

Keywords: Age structure, dry matter allocation, life history, Sparaxis grandiflora spp. fimbriata, storage organ

• To whom correspondence should be addressed.

Introduction Material and Methods Sparaxis grandiflora spp. Jimbriata (Lam.) Goldbl., is a Study site and sampling methods connous member of the family lridaceae, endemic to the The study site was located near the Blackheath industrial south-western Cape (Goldblatt 1969) where it is confined to area (34°62'E and 18°43'S). This site represents the fairly damp clay flats and hills in the Cape Town vicinity remains of a natural habitat of Sparaxis grandiflora spp. (Goldblatt 1%9). It is a synanthous geophyte (sensu Dafni Jimbriata disrupted as a result of sprawling industrial et at. 1981 b) which possesses an annual storage organ. development in the area. The site is flat and waterlogged In this region with its strongly seasonal rainfall pattern during the major part of the winter growing season. combined with periodic drought and soils of low nutrient A completely randomized block of 144 m2 was demarc­ status (Witkowski 1989; Witkowski et at. 1990), a plant is ated in the study area and this was subdivided into 12 likely to be at a competitive advantage if it possesses an quadrats of 12 m2 each. Sampling was done monthly using underground storage organ which pennits it to survive un­ random pennutations to select the quadrat (Moses & Oak­ favourable conditions, and allows the plant to carry a sub­ ford 1%3). Five sods (20 x 24 cm width and 10 - 15 cm stantial fraction of its nutrient resource from one growing deep) were removed each month from the quadrat (by de­ season to the next. structive digging) and the plants in the sods were carefully Knowledge of longevity, population age structure, and separated in the laboratory for analysis. Five soil samples generation span are lacking for most geophytic species in were also taken each month for the detennination of the soil their natural fynbos habitat. This paper examines life fonn, moisture content, which was measured as a percentage of phenology and reproductive behaviour in relation to factors the soil dry mass (dried at 80°C until a constant mass was such as depth of burial, size and age of the conns, the reached). number of leaves produced per conn, and the number of individuals per age class. This investigation is part of an Plant material and harvesting of plants in-depth long-tenn study into strategies adopted by the A total of 2880 plants were excavated from the site during geophytic guild in the fynbos biome (Ruiters 1990). the study period over the 1988/89 season. Records were S.AfrJ.Bot., 1992,58(3) 183

"--'1 . '- ~

:;~r:;··1:h;.~; CO tCR CR ~

Winter/Spring Summer/Autumn '

Figure 1 Ontogenic patterns in development and burial of the underground storage organ in Sparaxis grandiflora spp. fvnbriala (Lam.) Goldbl. The timing of the principal events is related to a calender year. E, endosperm; cr, cotyledon; R, radicle; L, seedling leaf; CR, contractile root; P, plumule; CO, corm.

taken of: leaf number, number of corm lets produced per Results and DiSCUSSion plant, number of flowers and fruits produced per plant, Ontogeny diameter and depth of burial of corms, and the number of This species displayed a form of germination in which the tunics around the corms. This latter feature was of great cotyledon extended greatly in length both above and below relevance to the ageing of plants because these tunics are the ground after germination (Figure 1). Thereafter, the highly resistant to decay (Dixon & Pate 1978), and the plumule was pulled underground by the combined action of parent corm is replaced in situ by the daughter corm, which the downward growth of the basal region of the cotyledon suggested that tunic counts could be used as an indicator of and the contractile activity of the primary root. The latter plant age. Hence, with time a series of concentric layers was the only root to form in the first season of growth. One build up around a corm, the outermost of these layers being to three leaves were formed in the first season. By the time the remnant from the season of recruitment as a new of dormancy (mid-November) at or near the end of the first daughter corm (Dixon & Pate 1978; Pate & Dixon 1982) growing season, stem tissue had swollen below the apex to and the innermost layer being that of the previous season's form a diminutive corm and all the other parts of the seed­ corm in which the present season's replacement (daughter) ling, including the contractile root, had withered and died. corm is currently developing. Tunic number was taken as a direct measurement of plant age in years. Year one in the Seasonal phenology and biomass accumulation life of a plant was defined as the 12-month period following Seasonal changes in distribution of dry matter are illustrated its appearance as the corm from seed or cormlet. The count in Figure 2. Corm dormancy spanned the five hot dry of 18 tunics, the result of in situ replacement of the parent months after which the parent corm broke dormancy (during corm, indicated the resistance of these dead layers of tissue late March), probably coincident with lower soil and air to decay, and how long certain plants can continue to re­ temperatures during autumn but before the soil was fully place their corms under natural conditions. It seemed drenched by the autumn rains (Figures 3 & 4). reasonable to assume, therefore, when examining popula­ The leaves grew to the soil surface in early April and had tions whose members proved to be much smaller and possessed fewer tunics than mentioned above, that errors 16,------.------. due to loss of outer tunics by decay would be relatively l1li Fruit 14 DFlowera unimportant. Goldblatt (1982, 1985) has found for Geisso­ Il)Ilnll .•talk rhiza and Hesperantha, genera related to Sparaxis, that Imleaves 12 corm tunic layers are produced annually and are also highly IIIIDl!lughler corm ~Par8nt corm resistant to decay. .Roots

The sampled plants were separated into their constituent en ~ 8 organs, viz., parent corms, replacement or daughter corms, E o cormlets, roots, stems, leaves, flowers, and fruit and seed. m 6 The respective plant organs of each sod were pooled and dry weight was then determined. 4

The above factors were analysed so that the ontogeny, 2 phenology of growth and reproduction, and age structure of the species could be determined. The monthly collections o were carried out to follow seasonal changes in dry matter Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Months content of plant parts and to examine the filling of the season's new corms with dry matter and mineral elements. Figure 2 Seasonal dry mass (g) changes in plant parts of a The results for the mineral element allocation are given by Sparaxis grandiflora spp. fimbriala popUlation during a calender Ruiters (1990) and Ruiters et al. (submitted). year in a natural habitat at Blackheath. 184 S.-Afr.Tydskr.Plantk., 1992,58(3)

I Summer I ~1------~A~ut~um~nNV~in~te-r------~ ~------S~p-r~in-g------~ Summer

Figure 3 Seasonal phenology and growth cycle of a Sparaxis grandiflora spp. fimbriala (Lam.) Goldbl. population in a natural habitat at Blackheath. C, corm; DC, daughter corm; PC, parent corm; IS, inflorescence stem; Fl, flowers; Fr, fruit; L, leaves; R, adventitious roots; T, tunics.

30.------, plants. The leaves remained green and active until senescence in October (Figures 2 & 3). Allocation of dry 25 matter to the daughter corm and reproductive structures was

~ the major assignment from June to November. The newly­ ~ formed photosynthate must have been the major source of -c 20 .2 co filling the daughter corm, since the above-ground vegetative o o and reproductive parts did not show net losses of dry matter ~t5 il.. over the period when maximum accumulation of dry matter ·0 was made by the daughter corm. The parent corm initially E 10 showed a slow decline in dry matter during the second half of the growing season, dropped abruptly from September to 5 October and was totally exhausted by November. By August the dry matter of the daughter corm had become greater than the initial dry matter of the parent corm. Maximum dry Jan Feb Mar Apr May Jun Jui Aug Sep Oct Nov Dec matter allocation in the plant was achieved during Sep­ Months tember. Figure 4 Mean monthly (:tSE) soil moisture content (%) for This species, which flowered from August to September, the study site at Blackheath. partitioned some of its dry matter to flower and fruit formation while still in the process of filling the daughter corm. Thus, competition for photosynthate exists between expanded above the soil surface by mid-April. Development young fruit and the daughter corm in October and it might of new adventitious roots took place in association with leaf well be an important factor affecting seed production and development. Establishment of the above-ground plant parts the underground reserves of the species (pate & Dixon in the early season of growth involved expenditure of dry 1982). matter from the parent corm (Figures 2 & 3). By June a sub­ The above-ground vegetative and reproductive plant parts stantial portion of the parent corm dry matter had been had senesced by mid-November and by the end of Decem­ exhausted. This species initially produced two to three ber the only living plant part that remained was the daughter leaves during autumn, but as the season progressed more corm which would become the parent corm of the new leaves were produced up to a maximum of twelve in some season. On entering dormancy the combined season's pro- S.Afrl.Bot.,1992,58(3) 185

duction of dry matter for seed and daughter corm accounted mean corm diameter with only minor fluctuations in the 9 - for about 78.0% of the total dry matter production. Only 12-year age classes. small amounts were incorporated into seed production From the data comparing corm diameter with depth of (2.3%) in comparison with the amount laid down in the burial the largest corm diameter was obtained in the 2 - 4- storage organ (75.7%) (Figure 2). Material for the filling of cm zone of the soil profile. The smallest corm diameter, as the seeds might have come from the vegetative parts and expected, was obtained in the 0 - 1.5-cm zone. Considerable flowers since they had lost substantially greater amounts of variation was evident for the 5 - 8-cm zone, which indicated dry matter during the senescence period than was required that corm diameter of reproductively mature individuals was for the growth of the seeds (Figures 2 & 3). Thus this not positively correlated with depth of burial in the soil species employed a highly efficient mechanism of retrieving profile (y = 0.930 - 0.0313x, r = -0.0313, t(34) = -0.56, dry matter for its daughter corm and seeds. The reproductive p = 0.576), possibly due to the burrowing of mole-rats and structures accounted for a small amount of production, with gerbils. The number of leaves produced per corm increased maximum fruit production of 0.43 g dry mass per plant progressively with corm diameter. A maximum of twelve during October, while production of mature seed averaged leaves was found in corms with a diameter greater than 14.29 ± 0.69 seeds per fruit. Total reproductive effort, 1 cm. including biomass allocation to the inflorescence stalk, floral The production of new individuals in S. grandiflora spp. parts and fruit development, was between 19.72 and 42.94% Jimbriata occurred asexually through the vegetative of the observed net production between September and Oc­ formation of cormlets and sexually through production of tober. Biomass in mature seeds accounted for 4.92 - 11.24% seeds. Vegetative reproduction only occurs in flowering of net production, and 25.9 - 38.4% of the reproductive individuals as cormlets are only formed in the leafaxils of effort. the flowering stem. In other species there is apparently a

Population parameters A series of instructive relationships emerged between depth of burial of the plant's corm, size of daughter corm, number of leaves produced, age of plants and flowering of plants in the distinct age classes in the year of study. The presumed maximum age using tunic counts was found to be 18 years (Figure 5). The shape of the age dis­ tribution histogram suggests that 18 years may be near the maximum age attained by individuals on the Blackheath study site. The histogram shows strong positive skewness (to the left; v:x+378 transformed) which indicates that the bulk of the population (86.1 %) was found in the 1 - 7 -year age classes. Most of the variation in size of the adjacent year classes could be due to variations in initial germination and 1 2 3 4 5 6 7 8 9 101112131415161718 seedling survival successes in different years (Kerster 1%8). Presumed oge (yeors) However, longer term studies are in progress which will enable the determination of the characteristics of population Figure 5 Frequency distribution (N, yIX+378 transformed) of size, age specific fecundity and survival, and density. the number of individuals per age class in a Spar axis grandiflora From the data on presumed age versus depth of corms in spp.fimbriala (Lam.) Goldbl. popUlation. the soil profile a progressive increase in the mean depth per age class for the first seven years was found. A constant 1.2,------. distribution for depth of burial was attained for the 8 - 16- year age classes (4.08 ± 0.51 cm). About one-quarter (27.8%) of the population's plants were found to be located within this range. The deepest depth of burial was found in 10 .8 the 18-year age class. There was some evidence that older .,~ plants were located in the 4 - 8-cm zone range, but it was Q; E 0.6 o found that age has apparently little effect on the overall i:5 mean depth of corm burial, the final depth on average being E 50.4 reached after eight years. The organic layer of the soil o profile extends to 3 cm and 80% of the sampled plants were found to be located within this zone. Since the species does 0 .2 not possess an extensive root system it is important for the plant to exploit this relatively nutrient-rich organic layer. Corm diameter for the various age classes was determined 1 2 3 4 5 6 7 8 9 1011 12 13 14 15 16 17 18 Age in Yeors for the population and is presented in Figure 6. A relatively steep increase in corm diameter was found for the first four Figure 6 Corm diameter (cm) for the various age classes of a years. Thereafter, the population exhibits uniformity in the Sparaxis grandiflora spp.fimbriata (Lam.) Goldbl. popUlation. 186 S.-Afr.Tydskr.Plantk., 1992, 58(3)

density-related influence on the relationship between vege­ tative and sexual reproduction (Muller 1979). In dewberries, Rubus hispidus and R. trivia lis, under conditions of high population density the proportion of biomass allocated to reproduction is increased in favour of sexual reproduction (Abrahamson 1975). S. grandiflora spp.fimbriata at this site 2), '"c 60 has a high density (84.2 m- and the allocation in terms of ·c., 30 biomass to corm lets was insignificant in this study, but this 0 does not necessarily indicate that vegetative reproduction is G: ~ 40 unimportant (Goldblatt 1969). It was found that 77.4% of plants with a corm diameter greater than 0.78 cm and a bio­ mass greater than 0.37 g, sampled during the flowering 20 season, did flower (n = 199) and that 86.6% of plants with a corm diameter below 0.76 cm (0.37 g d.m.) did not flower (n = 238). Furthermore, plants younger than four years did 2 3 4 5 6 7 8 9 10 not flower (Figure 7, arcsiny'x transformed). This suggested Age in years that this synanthous geophytic species requires a 'minimum Figure 7 Percentage of individuals (arcsinVX transformed) of critical' corm biomass and size (Rees 1969, 1972; Dafni et each age class that flowered in a Sparaxis grandiflora spp. al. 1981a, b) before flowering can occur. The data in Figure jimbriata (Lam.) Goldb!. population (n = 437). 7 only show percentage flowering up to ten years of age as no individuals older than ten years were present in the sample. This does not imply that individuals older than ten Acknowledgements years did not flower, but their contribution to recruitment We thank Mr Neethling for permission to work on his would in any case be low, as there are very few individuals property, W.J. Bond for comments on an earlier draft and in the age groups older than ten years (Figure 5). Further the three anonymous referees for their comments, C.T. investigations on the population profile are continuing and Johnson for the recruitment of funds, Ms H. Taylor for the should be able to address this facet. Age and biomass are drawings and J. Aalbers for the translation of the abstract. equally important predictors of flowering status, but for a This project was funded by the De Beers Chairman's Fund. number of species total biomass has been emphasized as a better predictor for flowering than age (Pritts & Hancock References 1983). This is to be expected, as biomass would be a better ABRAHAMSON, W.G. 1975. Reproductive strategies in indication of physiological rather than chronological age, as dewberries. Ecology 56: 721 - 726. is the plastochron index (Erikson & Michelini 1957; Lamou­ DAFNI, A., SHMIDA, A. & A VIS HAl, M. 1981a. Leafless reaux et al. 1978). autumnal-flowering geophytes in the Mediterranean region - Phytogeographical, ecological and evolutionary aspects. Pl. Conclusions Syst. Evol. 137: 181 - 193. The phenology, demography and utilization of resources for DAFNI, A., COHEN, D. & NOY-MEIR, I. 1981b. Life-cycle variation in geophytes. Ann. Mo . bot. Gdn. 68: 652 - 660. reproduction in S. grandiflora spp. Jimbriata resemble those DIXON, K.W. & PATE, 1.S. 1978. Phenology, morphology and of many annual synanthous geophytic species, with charac­ reproductive biology of the tuberous sundew, Drosera erythro­ teristics such as annual die-back of above-ground herb­ rhiza Lind!. Aust. l. Bot. 26: 441 - 454. aceous plant parts and that a relatively large proportion of ERIKSON, R.O. & MICHEUNI, F.l. 1957. The plastochron the annual biomass production is invested in filling the index. Am. l . Bot. 44: 297 - 305. storage organ and in reproductive effort. The development GOLDBLATT, P. 1969. The Sparaxis. ll. S. Afr. Bot. 35: of an underground annual storage organ coupled with the 219 - 252. annual die-back of above-ground herbaceous plant parts in GOLDBLATT, P. 1982. Corm morphology in Hesperantha (lrida­ S. grandiflora spp. Jimbriata can be regarded as an adaptive ceae; Ixioidaceae) and a proposed infrageneric . Ann. strategy to retain nutrients and avoid drought and thus cope Mo. bot. Gdn. 69: 370 - 378. with the low-nutrient status and climatic conditions of the GOLDBLATT, P. 1985. Systematics of the South African genus mediterranean-type ecosystem in the south-western Cape. Geissorhiza (lridaceae - Ixioideae). Ann. Mo. bot. Gdn. 72: The counting of corm tunics appears to be a reliable tool 277 - 447. for establishing the age of the species. Sexual reproduction KERSTER, H. W. 1968. Population age structures in the prairie starts at four years and one hundred percent flowering forb, Liatris aspera. Bioscience 18: 430 - 432. LAMOUREAUX, R.T., CHANEY, W.R. & BROWN, K.M. 1978. occurs at nine years. Population parameters such as corm The plastochron index: a review after two decades of use. Am. size, corm depth, age and biomass were found to be equally l . Bot. 65: 586 - 593. important predictors of flowering status. The effect of MOSES, L.E. & OAKFORD, R.Y. 1963. Tables of Random density on relative proportion of modes of reproduction for Permutations. Stanford University Press, Stanford, California. this species seems inappropriate since asexual reproduction MULLER, R.N. 1979. Biomass allocation and reproduction in in this species is dependent on sexual reproduction. The Erythronium albidum. Bull. Torrey bot. Club 106: 276 - 283. acquisition of a 'minimum critical biomass' for the corm is PATE, 1.S. & DIXON, K.W. 1982. Tuberous, Cormous and an important prerequisite for flowering which appears to Bulbous Plants. University of Western Australian Press, Ned­ take place when plants reach the age of four years. lands, Western Australia. S.AfrJ.Bot.,1992,58(3) 187

PRITIS, M.P. & HANCOCK, J.F. 1983. Seasonal and lifetime (submitted). Seasonal biomass and resource allocation patterns allocation pattern of the woody goldenrod, Solidago pauciflos­ for geophytic species on lowland coastal fynbos sites in the culosa Michaux. (Compositae).Am. 1. Bot. 70: 216 - 22l. southwestern Cape, South Africa. II. In a population of a REES, A.R. 1969. Effects of bulb size on the growth of tulips. perennial cormous geophyte [Sparaxis grandiflora subspecies Ann. Bot. 33: 133 - 142. ftmbriata (Lam.) Goldbl.]. Ann. Bot. REES, A.R. 1972. The Growth of Bulbs. Academic Press, WITKOWSKI, E.T.F. 1989. Response to nutrient additions by the London. plant growth forms of sand-plain lowland fynbos, South RUITERS, C. 1990. Population and Physiological Studies of Africa. Vegetatio 79: 89 - 97. Selected Geophytic Species of Lowland Coastal Fynbos in the WITKOWSKI, E.T.F., MITCHELL, D.T. & STOCK, W.O. Southwestern Cape. Unpublished M.Sc. thesis, University of 1990. Response of a Cape fynbos ecosystem to nutrient addi­ the Western Cape, Bellville, South Africa. tions: nutrient dynamics in fertilized soils. Acta Oecologica 11: RUITERS, C., McKENZIE, B., AALBERS, J. & RAITI, L.M. 165 - 179.