S.Afr.J.Bot., 1991,57(6): 319 - 324 319

Growth of calcicole and calcifuge Agulhas Plain Proteaceae on contrasting soil types, under glasshouse conditions

I.P. Newton,* R.M. Cowling and O.A.M. Lewis Botany Department, University of Cape Town, Private Bag, Rondebosch, 7700 Republic of South Africa

'Present address: Fitzpatrick Institute, University of Cape Town, Private Bag, Rondebosch, 7700 Republic of South Africa

Accepted 16 August 1991

Seven species of Proteaceae - Protea compacta, P. obtusifolia, Leucadendron xanthoconus, Ld. meri­dianum, Leucospermum cordifolium, Ls. truncatum and two populations of Aulax umbellata - from the Agulhas Plain, south-western Cape, were grown in acid sand and neutral limestone sand in pots in a greenhouse. Heights were measured monthly, and after 400 days the plants were harvested and the dry weights of shoots and roots determined. Species normally growing on acid sands showed signs of chlorosis and necrosis when grown on limestone soils, and weights at harvesting were significantly less than for the same species grown on acid sand. Species normally growing on limestone soils were not significantly different in height or weight when grown on either soil type. Mean species root weight in all but one case was more for plants grown on acid sands than on limestone soils. It is suggested that soil type plays the major role in the failure of calcifuge species to establish themselves on limestone soils, but that competition is probably more important in the reverse case. Some suggestions for future research into the physiological tolerances of conspecifics under controlled conditions and field conditions are made.

Sewe spesies van Proteaceae - Protea compacta, P. obtusifolia, Leucadendron xanthoconus, Ld. meri­dianum, Leucospermum cordifolium, Ls. truncatum en twee populasies van Aulax umbellata - van die Agulhas Plein, suid-westelike Kaapprovinsie, is in suurgrond en in neutrale kalksteengrond in plastiekhouers in 'n glashuis gekweek. Planthoogtes is maandeliks gemeet, en die plante is na 400 dae gebes en die massa van die stingels en wortels is bepaal. Spesies wat normaalweg in suurgrond groei, het tekens van chlorose en nekrose getoon as hulle in neutrale grond gegroei het, en die oesmassa was heelwat laer as die van dieselfde spesies wat op suurgrond gegroei het. Spesies wat normaalweg in neutrale grond groei, het geen noemenswaardige verskille in massa of hoogte getoon wanneer hulle op die verskillende grondsoorte gegroei het nie. Die gemiddelde wortelmassas was, behalwe in een geval, deurgaans groter vir plante wat op suurgrond gegroei het as vir plante wat op neutrale grond gegroei het. Die implikasies van die bepaalde verspreiding van hierdie spesies word kortliks bespreek. Daar word voorgestel dat grondtipe die hoofsaaklike rede is waarom kalksku spesies nie op neutrale kalksteengrond groei nie, maar in die teenoorgestelde geval is kompetisie 'n belangriker faktor. Voorstelle word gemaak vir toekomstige navorsing oor fisiologiese toleransies van soortgelyke spesies onder gekontroleerde en veldtoestande.

Keywords: Calcicole, calcifuge, growth, Proteaceae.

Introduction The coastal forelands of the southern Cape include extensive areas of limestones of the Bredasdorp formation, which date from the Mio--Pliocene era and support fynbos communities dominated by shrubs (l.0 - 2.0 m) belonging to the Protea­ceae (Cowling, Campbell, Mustart, McDonald, Jarman & Moll 1988; Rebelo, Cowling, Campbell & Meadows 1991). Most of these proteoid species are calcicoles (confined to soils with a high calcium content), and many are local endemics (Dahlgren 1963; Nordenstam 1969; Cowling, Holmes & Rebelo, in press). Immediately adjacent to these limestone outcrops are leached acid sands which support a structurally similar, but floristically distinct, proteoid fynbos (Thwaites & Cowling 1988; Cowling et al. 1988; Rebelo et al . 1991). These juxtaposed communities share very few species (Cowling 1990).

What limits the distribution of calcicole and calcifuge (confined to low-calcium soils) species in the southern Cape fynbos? Despite the potential these communities provide for an experimental analysis of the calcicole--calcifuge question, no research has been done. Studies in the northern hemi­sphere have indicated that in natural, as opposed to agri-

cultural, situations, factors such as competItIon (Snaydon 1962), particularly in the seedling stage (Rorison 1960), play an important role in species distribution. Grubb, Green and Merrifield (1969) showed that seeds of both calcicole and calcifuge species on an English chalk heath germinated successfully, and both developed healthy rooting systems, when the soil pH was between 5 and 6.

Interactions between inorganic elements, pH, organics, saturation and aeration levels of the soil all play a part in determining how well, or even whether a plant can grow in a particular soil type. It is not our intention to even attempt to summarize the literature, but a few major generalizations follow. Increasing solubility of certain metallic ions as the soil solution pH drops, potentially reaching levels toxic to calcicole plants, can be a major reason for calcicole species failing to grow on ac idic soils (Clymo 1962). Aluminium is particularly important because as little as 1 p.p.m. can have a detrimental effect on root growth at pH 4.2. (Clymo 1962, Rorison 1960). This toxicity at low pH is one of the reasons for the liming of agricultural soils (Bartlett & Riego 1972). Low calcium levels in acid soils may also cause growth fail ­ure, largely because calcium is involved in the uptake of

320

other ions (Amon, Fratzke & Johnson 1942). An increase in pH can lead to deficiencies in calcifuge plants of metaIlic elements such as iron (Holmes 1960; Lucas & Davis 1961), manganese (Gilbert & Mclean 1928), boron and zinc (Lucas & Davis 1961), and magnesium and potassium (Snaydon 1962). The base ratios of potassium and sodium to calcium and magnesium are also considered to have an effect (pearsall & Wray 1927).

This paper reports on a preliminary study of the calci­cole-calcifuge effect on fynbos Proteaceae. The aim of the study was to determine whether there were differences in the relative growth rates of seedlings of acid and limestone species when grown under glasshouse conditions on their native and on their neighbouring soil types. Even smaIl differences in relative growth rates may translate into significant competitive effects (Givnish 1986). Although glasshouse trials are limited in their application to the field, our study does provide an indication of the edaphic effect on the distribution of fynbos Proteaceae.

Methods Seeds of seven Proteaceae species growing on the Agulhas Plain on acid sand or on limestone-derived alkaline-neutral sand were collected in January and February 1988. The species concerned were (calcifuge species first for each genus): Protea compacta R.Br. and P. obtusifolia Bucek ex Meisn., Leucadendron xanthoconus (0. Kuntze) K. Schum. and Ld. meridianum Williams, Leucospermum cordifolium (Salisb ex Knight) Fourc. and Ls. truncatum (Buek ex Meisn.) Rourke, and two populations of Aulax umbel/ata Berg., one growing on acid sand (the normal habitat) and the other from a smaIl population on limestone. Soils from each site were coIlected at the same time. In October 1988 seeds of two populations of P. obtusifolia were collected. One population was from limestone (the normal habitat), the other from nearby acid sand. Soil from the site of the acid sand population was collected.

Seeds and soils were retumed to the laboratory. To en­courage germination, the Protea and Aulax seeds were treated according to Vogts (1982); the Leucospermum seeds were soaked in concentrated sulphuric acid for eight minutes and then left to soak in a one percent hydrogen peroxide solution for 24 h (modified from Vogts 1982). Leucaden­dron seeds were untreated. All the seeds were then lightly dusted with Thiram fungicide powder, and placed on damp filter paper in closed Petri dishes. These were placed in a constant-environment chamber set to 10 h light at 20°C and 14 h dark at 9°C. Most seeds germinated after two to three weeks in these conditions, but the Leucospermum seeds took longer and many failed to germinate.

Once the radicle began developing the seeds were planted out into 20 cm diameter pots, three plants per pot. For each species there were three pots each of an acid and a lime­stone soil. Each species was planted into the soil type from which it had been collected, and into the soil type of its congener. The pots were arranged in a block of 4 x 12 pots. The plants were watered three times per week with between 150 and 500 ml of de-ionized water, depending on tempera­ture, plant size and soil type, thus attempting to keep water­related stresses to a minimum. Each month the height of each individual plant was measured (top of cotyledon to the tip of the top leaves folded upwards), and the pots were

S.-Afr.Tydskr.Plantk., 1991,57(6)

moved with respect to each other to compensate for one end receiving more sun. The age of each plant was calculated from the date that the cotyledons were first fully expanded. The plants were grown in a lean-to greenhouse on the east side of the Botany Department. The south end of the green­house received some afternoon sun in summer. The green­house had thermostatically controIled heating and tempera­tures ranged from 20 to 40°C with 50 to 90% humidity.

Plants were harvested in June 1989, at which stage the plants were around 400 days old. Dry weights of roots and shoots were measured.

Seedlings of the two HageJkraal populations (acid sand and limestone) of P. obtusifolia were planted in 20 cm diameter pots (three per pot) containing acid sand. Condi­tions of growth of these plants were the same as for the above plants, except that the pots were arranged in a block of 3 x 7, edaphic populations alternating. Plants were harvested in May 1990 when they were around 420 days old. Fresh and dry weights of roots and shoots were de­termined. As above, plants dying before the harvest were excluded from the analysis.

Results All the soils had low levels of nutrients, particularly total nitrogen and available phosphorus (Table 1). Limestone­derived sands had higher pH, total nitrogen, available phosphorus, available calcium and organic carbon than the acid sands. However, a big variation in nutrient levels was recorded between limestone sites, and a good deal more soil analyses are required.

There were no major differences in the growth rates (seedling height) of seedlings of calcicole species on contrasting soil types (Figure 1). The growth rate of the calcicole Aulax population was greater on acid sand than on limestone, although not significantly so (DN = 0.385; P = 0.291; Komolgorov-Smimoff two-sample test) (Siegel 1956). There were no obvious differences in the appearance of calcicole seedlings growing on the different soil types. There were also no significant differences in the dry weights of whole plants, roots or shoots for calcicole plants growing on the different soil types (Table 2). With the exception of Aulax, root: shoot ratios were consistently higher for limestone-grown seedlings. Although the calcicole Aulax seedlings grew taller on acid sand, the limestone-grown seedlings weighed more, largely due to the more extensive rooting system.

Growth rates of seedlings of all calcifuge species were lower on limestone-derived than on acid sand, although only significantly so for Aulax (DN = 0.571; P < 0.05). The growth curve of the calcifuge Aulax on limestone differed significantly from those of the calcicole population growing on both acid sand (DN = 0.57; P < 0.05) and on limestone (DN = 0.538; P < 0.05). The calcicole Aulax seedlings grown on acid sand had lower dry weights than calcifuge seedlings on acid sand, although not significantly so (P = 0.2167). The calcifuge Aulax seedlings on limestone had a significantly lower dry weight than those of the calcicole population growing both on limestone (P < 0.(05) and on acid (P < 0.0001) sand.

There were numerous visual differences between seed­lings of calcifuge species growing on the two soil types.

S.Afr.J.Bot., 1991, 57(6) 321

Table 1 Study species and soil data from the Agulhas Plain

Soil properties'

pHb Total NO)" NH4d P K Na Ca Mg Fe Al CEC Organ. Base Site Species N (%) (c. acid) (EDTA) me% C (%) ratio

Soetanysberg

TMG sandstone Leucospermum cordifolium 5.8 0.018 0.65 0.62 1.6 16.1 6.4 2.9 19.4 0.8 0.75 0.45 1.90

Colluvial Au/ax umbel/ala, 4.7 0.013 0.33 0.29 <1 2.0 13.8 5.2 1.8 12.0 0.3 0.55 0.69 2.25 acid sand Leucospermum Iruncalum,

Prolea compacla

Heuningrug

Bredasdorp Leucadendron meridianum, 6.9 0.076 2.96 0.63 2 5.9 85.1 105.8 7.4 74.8 0.5 5.22 2.72 0.80 limestone Leucospermum Iruncatum,

Protea obtusifolia

Hagelkraal

Bredasdorp Au/ax umbel/ata, 6.8 0.029 0.10 0.15 3 0.8 32.2 68.6 1.5 14.3 0.5 2.79 0.86 0.47 limestone Protea obtusifolia 6.7 0.170 9 75.9 195.4 9.3 13.95

Colluvial Protea obtusifolia 4.4 0.017 0.2 34.5 10.4 3.1 0.65 acid sand

• All values are micrograms per gram dry weight of soil, unless specified otherwise. All analyses were done by the Eisenberg Agricultural College, except for those described below.

b 0.01 M Ca02 (Schofield & Taylor, 1955). " Szechrome NAS method (Stock 1983). d Indo-Phenol Blue method (Stock 1983).

Table 2 Dry weights of ca. 400-day-old pot-grown calcicole and calcifuge Proteaceae from the Agulhas Plaina

Dry weight (g) Root: shoot

Speciesb Soil type" Plant Shoot Root ratio N

Aulax umbel/atab Acid sand 2.27 :!: 0.29 1.73 :!: 0.20 0.54 :!: 0.10 1:3.2 7 Limestone 0.24 :!: 0.05 0.21 :!: 0.04 0.03 :!: 0.01 1 :11.6 7

*** *** *** Aulax umbel/ala Acid sand 1.98 :!: 0.18 1.60 :!: 0.17 0.38 :!: 0.06 1:4.2 6

Limestone 2.11 :!: 0.49 1.43 :!: 0.28 0.69 :!: 0.21 1:2.1 8 N.S. N.S. N.S.

Leucadendron meridianum Acid sand 3.44 :!: 0.54 2.42 :!: 0.36 1.02 :!: 0.20 1:2.3 9 Limestone 4.27 :!: 0.73 3.45 :!: 0.59 0.82 :!: 0.15 1:4.2 9

N.S. N.S. N.S.

Leucadendron xanthoconusb Acid sand 3.49 :!: 0.24 2.59 :!: 0.20 0.90 :!: 0.06 1:2.9 9 Limestone 1.84 :!: 0.34 1.36 :!: 0.25 0.48 :!: 0.09 1:2.8 9

** ** ** Leucospermum cordifoliumb Acid sand 8.07 :!: 2.81 5.53 :!: 1.64 2.54 :!: 1.19 1:2.2 4

Limestone 4.03 :!: 0.71 3.24 :!: 0.57 0.79 :!: 0.15 1:4.1 9 N.S. N.S. *

Leucospermum truncatum Acid sand 5.66 :!: 1.59 3.30 :!: 1.03 2.36 :!: 0.59 1:1.4 5 Limestone 5.52 :!: 0.77 4.50 :!: 0.65 1.03 :!: 0.13 1:4.4 4

N.S. N.S. N.S .

Protea compacta b Acid sand 6.45 :!: 0.70 4.58 :!: 0.59 1.87 :!: 0.18 1:2.4 8 Limestone 3.25 :!: 0.60 2.73 :!: 0.44 0.52 :!: 0.16 1:5.3 8

** * *** Protea obtusifolia Acid sand 1.45 :!: 0.20 1.10:!: 0.13 0.35 :!: 0.07 1 :3.1 9

Limestone 1.63 :!: 0.12 1.44 :!: 0.10 0.19 :!: 0.03 1:5.3 8 N.S . N.S. N.S.

• Plants were grown under glasshouse conditions in acid and limestone-derived sands. Data are mean:!: s.e. Significance (t-test): ***: P < 0.001; **: P < 0.01; *: P < 0.05; N.S.: Not significant.

b Species marked with a superscript are calcifuge, non-marked species are calcicole. " Limestone is from Heuningrug, except that for Aulax which is from Hagelkraal. Acid sand is colluvial sand from Soetanysberg,

except that for Leucospermum which is T.M.G.-derived sandstone from Soetanysberg.

Death of growing tip and leaf necrosis were observed in A. umbellata and P. compacta growing in limestone-derived soil. Both Ls. cordifolium and P. compacta seedlings show-

ed obvious signs of chlorosis in limestone-derived sand, although none of the Ls. cordifolium seedlings died on this soil. Five of the nine Ls. cordifolium seedlings established

322

E E -

CALCICOLE CALCIFUGE Aulax umbellata Aulax umbellata

200 ON = .385 200 l ON = .571 I

P = .291 , 1 ( ln , .... ( .· 1 ·

I , P = .021

! I

I 100 ~ ./"

~ .... . ·r· ""1" " T " r "

100

00 100 200 300 400 0 1

0 100 200 300 400

Leucadendron meridianum Leucadendron xanthoconus

400 f ON = .143 , I I 1 400 l DN = .357

200 r . < I 200 r . 300 fLP = 1 0 < 1 1 I 1 300 ~ P = .334

1 00 [ , . ' 1 1 00 •

00 1 00 200 300 400 0 0~~;:;:---;C;:;:-~30!cOo--"740!cO~

400

300

Leucospermum truncatum Leucospermum cordifolium

DN = .23

P = .999

500 F 400 r DN = .5

300 r P = .06

200 l 200 r .,' 100 l .. , I 100 r • . <. ,

o 0!:-=~1 0=-=0"--2=0=-=0"--3=0'-=0-4""'0~0-' 0 0 1 00 200 300 400

Protea obtusifolia Prate a compacta

::: r ,~": ::: f r / :N==1~~ i 100 t P = .999

o l . . I ~I _~,.--~,.--~_~,....J o 100 200 300 400 00 100 200 300 400

., ... , ... , " "{" """'I

ON = .308

Age (Days) Figure 1 Growth of seedlings of calcicole and calcifuge Protea­ceae from the Agulhas Plain, in limestone-derived (alkaline -dotted line) and sandstone-derived (acid - solid line) sands under glasshouse conditions. Vertical bars indicate one standard error (one side only to enhance clarity). DN: test statistic for the Kolmogorov-Smimoff two-sample tests (Siegel 1956). p: level

of significance.

on acid sand died of unknown causes. Ld. xanthoconus seedlings showed only slight chlorosis, but branching was minimal compared with those growing on acid soil.

Dry weights of roots, shoots and whole plants were greater for all calcifuge species when the plant was grown in acid sand than when the plants were grown in limestone (Table 2). These results were significant for all but the shoot and whole plant weights of Ls. cordifolium, where results were probably affected by the reduced sample size. As with the calcicole species, the root: shoot ratios were highest in seedlings grown on limestone-derived sand.

Seedlings of the limestone population of P. obtusifolia had significantly lower plant, shoot and root dry weights (P < 0.1) when grown on acid sand than had seedlings of an acid-sand population grown in the same soil (Table 3).

Discussion A clear trend for the growth of seedlings from populations naturally growing off their normal parent material was evi­dent. Both the limestone population of A. umbellata and the acid-sand population of P. obtusifolia showed enhanced performance on these 'foreign' soils, relative to seedlings from populations growing on their normal substratum.

The growth of seedlings of calcicole Proteaceae, under glasshouse conditions, is basically the same on limestone as

S.-Afr.Tydskr.Plantk.,1991,57(6)

Table 3 Dry weights of ca. 400-day-old pot-grown seedlings of Protea obtusifolia a

Dry weight (g) Root : Shoot

Population N Shoot Root Plant ratio

Acid sand 28 1.67 ::!: 0.18 1.39 ::!: 0.13 3.06 ::!: 0.26 1 : 1.20 Limestone 27 1.31 ::!: 0.21 1.10 ::!: 0.11 2.50::!: 0.28 1 : 1.19

* ** ** • Plants were established from seeds collected from populations growing on

acid sand and limestone-derived sand. Plants were grown in acid sand under glasshouse conditions. Data are means ::!: s.e. Significance (I-test): *: P < 0.10; **: P < 0.05.

on acid sands. However, the same cannot be said about the calcifuge species. With the exception of Ld. xanthoconus, the calcifuge species displayed clear signs of stress on lime­stone sands, and root development was severely retarded. Chlorosis developed in all calcifuge species on limestone soils, although it was minimal in Ld. xanthoconus. Irrespec­tive of preferred soil type, and with the exception of the calcicole Aulax, plants grown on acid soils had heavier roots, which was the result of the rooting system being more densely branched than in the limestone soils. This develop­ment was probably to increase nutrient uptake in the lower nutrient sandstone soils (Lamont 1982). Proteoid roots ap­peared only to develop in the case of P. obtusifolia, slightly more so in the sandstone soil. Reports by other workers (Jeffrey 1967; Lamont in Mooney, Kummerow, Moll, Orshan, Rutherford & Sommerville 1983) indicate that proteoid roots tend to develop to increase the uptake of phosphates. Why other species did not develop these roots is unclear.

The shoot: root ratio is greater in all the limestone-grown plants than it is in the acid sand-grown plants. The excep­tions to this are Ld. xanthoconus where the ratios are about the same, and the limestone A. wnbellata where the ratio on limestone is half that on acid sand. These ratios are quite high and correspond to 'nutrient rich' soil ratios (Mooney et at. 1983); the ratio nevertheless correlates with the report by Mooney et al. that the rooting system is larger in soils with low nutrient levels and that South African fynbos data are less clear, and sometimes contradict the general rule.

What factors could be responsible for these patterns in seedling performance on contrasting soil types? It is interesting that the iron (and organic carbon) content of the Heuningrug limestone-derived sand was four to six times higher than that of the acid sand. Chlorosis of calcifuge species growing in the acid soils may have developed as a result of the reduced iron availability often associated with soils of high pH (Lucas & Davis 1961). Although the phos­phorus levels in the limestone soils were slightly higher than in the acid sands, levels are still very low and are unlikely to affect intra-cellular availability of iron (Medappa & Dana 1970). It appears that the lower levels of calcium in the acid sands do not affect the growth of calcicole Proteaceae seedlings. The inhibition of root growth by metallic ions such as aluminium (Clymo 1962) in acidic soils does not seem to be involved as the levels are very low. Base ratios of [(K + Na)/(Ca + Mg)] are much greater in the sandstone soils, a factor which may be worth further investigation, as

S.Afr.J.Bot., 1991,57(6)

high values have been found to reduce plant growth (pearsall & Wray 1927). Controlled nutrient and pH growth experiments would go a long way towards establishing the physiological causes of this edaphic isolation.

Populations of the calcifuge A. umbel/ata (Rourke 1987) and the calcicole P. obtusifolia (Rourke 1980) were found growing off their preferred substratum. We have noted this for other species, namely P. susannae (calcifuge) on lime­stone and Ld. meridianum (calcicole) on sandstone. What is interesting is that the seedlings of these maverick popula­tions performed better on the unusual soil than did seedlings from the preferred substratum. This indicates that selection could be occurring for genotypes more tolerant of the unusual soil. After fire, wind could easily disperse seeds of serotinous Proteaceae, such as the experimental species, across soil boundaries (Bond 1988). Mosaics of juxtaposed acid and limestone sands, such as are found on the Agulhas Plain, would provide ideal conditions for this (Thwaites & Cowling 1988). Seedlings could establish themselves off their parent material, but would probably be out-competed by ecologically similar species growing on their natural soil type (Rorison 1960; Snaydon 1962). We have noted that P. obtusifolia establishes itself on acid sand only where P. compacta is absent. Similarly, one of the authors (R.M.C.) noted that P. susannae was found only on limestone outcrops that lacked P. obtusifolia. These problems are best resolved by field experiments in which both edaphic and competitive (intra- and inter-specific) factors are considered.

The results presented here suggest that calcicole species should establish themselves more readily on acid sands than vice versa. The successful establishment of a calcifuge population on limestone would involve intense selection, even in the absence of competitors. The establishment of calcicolous populations on the MiD-Pliocene limestones of the south-western and southern Cape represents a major selective force for speciation among the predominantly calcifuge fynbos flora (Cowling et al., in press). This has resulted in the evolution of many hundreds of calcicolous endemics confined to the Bredasdorp limestones (Dahlgren 1963; Nordenstam 1969).

We believe that future physiological studies on the calci­cole-calcifuge effect should concentrate on edaphic eco­types of predominantly calcicole (e.g. P . obtusifolia) or predominantly calcifuge (e.g. A. umbellata) species. The ad­vantage of this approach is that phylogenetic constraints need not be considered in interpreting the adaptive signifi­cance of physiological traits and processes (Wanntorp, Brooks, Nillson, Nylin, Ronquist & Steams 1989).

Acknowledgements We thank Mr 1. Bell and Mr J. Albertyn, respectively, for access to the Hagelkraal and Heuningrug sites. This study was funded by the FRD's Core Programme (R.M.C. and O.A.M.L.) and the South African Nature Foundation (R.M.C.).

References ARNON, D.I., FRATZKE, W.E. & JOHNSON, C.M. 1942.

Hydrogen ion concentration in relation to absorption of inor­ganic nutrients by higher plants. Plan! Physiol. 17: 515 - 524.

323

BARTLETT, R.J. & RIEGO, D.C. 1972. Toxicity of hydroxy aluminum in relation to pH and phosphorus. Soil Sci. 114(3): 194 - 200.

BOND, W.J. 1988. Proteas as 'tumbleseeds ': Wind dispersal through the air and over soil. S. Afr. 1. Bal. 54: 455 - 460.

CL YMO, R.S. 1962. An experimental approach to part of the calcicole problem. 1. Ecol. 50: 707 - 731.

COWLING, R.M. 1990. Diversity components in a species-rich area of the Cape Floristic Region. 1. Veg. Sci. 1: 669 - 710.

COWLING, R.M., CAMPBELL, B.M., MUSTART, P.J., MC­DONALD, D., JARMAN, ML & MOLL, E.J. 1988. Vegeta­tion classification in a floristically complex area: The Agulhas Plain. S. Afr. 1. Bal. 54: 290 - 300.

COWLING, R.M., HOLMES, P.M. & REBELO, A.G. (in press). Plant diversity and endemism. In: The ecology of fynbos. Nu­trients, fire and diversity, ed. Cowling, R.M. Oxford University Press, Cape Town.

DAHLGREN, R. 1963. Studies on Aspalalhus: phytogeographical aspects. Bal. NOl. 121 : 504 - 534.

GILBERT, B.E. & MCLEAN, F.T. 1928. A 'deficiency disease' : the lack of available manganese in a lime-induced chlorosis. Soil Sci. 26: 27 - 31.

GIVNISH, T.J. 1986. Introduction. In: On the economy of plant form and function, ed. Givnish, T.J. Cambridge University Press, Cambridge.

GRUBB, P.J., GREEN, H.E. & MERRIFIELD, R.C.J. 1969. The ecology of chalk heath: its relevance to the calcicole-<:a1cifuge and soil acidification problems. 1. Ecol. 57: 175 - 212.

HOLMES, R.S. 1960. Effect of phosphorous and pH on iron chlo­rosis of the Blueberry in water culture. Soil Sci. 90: 374 - 379.

JEFFREY, D.W. 1967. Phosphate nutrition of Australian heath plants 1. The importance of proteoid roots in Banksia (Protea­ceae) . Ausl. 1. Bal. 15: 403 - 411.

LAMONT, B. 1982. Mechanism for enhancing nutrient uptake in plants, with particular reference to mediterranean South Africa and Western Australia. Bal. Rev. 48: 597 - 689.

LUCAS, R.E. & DAVIS, J.F. 1961. Relationships between pH values of organic soils and availabilities of 12 plant nutrients. Soil Sci. 92: 177 - 182.

MEDAPPA, K.C. & DANA, M.N. 1970. The influence of pH, Ca, P and Fe on the growth and composition of the Cranberry plant. Soil Sci. 109: 250 - 253.

MOONEY, H.A., KUMMEROW, J., MOLL, EJ., ORSHAN, G., RUTHERFORD, M.C. & SOMMERVILLE, J .E.M. 1983. Plant form and function in relation to nutrient gradients. In: Mineral nutrients in mediterranean ecosystems, ed. Day, J.A. S. Afr. Nal. Sci. Prog. Rep. 71: 55 - 76.

NORDENSTAM, B. 1969. Phytogeography of the genus Euryops (Compositae) : a contribution to the phytogeography of south­ern Africa. Opera Bal. 23 : 1 - 77.

PEARSALL, W.H. & WRA y, E.M. 1927. The physiology and ecology of the calcifuge habit in Eriophorum angustifolium. 1. Ecol. 15: 1 - 32.

Rf:BELO; A.G., COWLING, R.M., CAMPBELL, B.M. & MEADOWS, M.E. 1991. Plant communities of the Riversdale Plain. S. Afr. 1. Bal . 57: 10 - 28.

RORISON, I.H. 1960. Some experimental aspects of the calci­cole-<:a1cifuge problem. I The effects of competition and mineral nutrition upon seedling growth in the field. 1. Ecol. 48: 585 - 599.

ROURKE, J.P. 1980. The Proteas of southern Africa. Purnell, Johannesburg.

ROURKE J.P. 1987. The inflorescence morphology and systemat­ics of Au/ax (Proteaceae) . S. Afr. 1. Bal. 53: 464 - 480.

SCHOFIELD, R.K . & TAYLOR, A.W. 1955. The measurement of soil pH. Soil Sci. Soc. Am. Proc. 19: 164 - 167.

324

SIEGEL, S. 1956. Non parametric statIstIcs for the behavioral sciences, pp. 127 - 136. McGraw-Hill, Tokyo.

SNA YOON R.W. 1962. The growth and competitive ability of contrasting natural populations of Trifoliwn repens L. on cal­careous and acid soils. 1. Ecol. 50: 439 - 447.

STOCK, W.D. 1983. An evaluation of some manual colorimetric methods for the determination of inorganic nitrogen in soil extracts. Commun. Soil Sci. Plant Ann. 14: 925 - 936.

S.-Afr.Tydskr.Plantk., 1991,57(6)

THWAITES, R.N. & COWLING, R.M. 1988. Soil-vegetation relationships on the Agulhas plain, South Africa. Catena 15: 333 - 345.

VOGTS, M. 1982. South African Proteaceae: Know them and grow them. C. Struil(, Cape Town.

WANNTORP, H.E., BROOKS, D.R., NILLSON, T., NYLIN, S., RONQUIST, F. & STEARNS, S.c. 1990. Phylogenetic ap­proaches in ecology. Oikos 57: 119 - 132.