A survey of the mycobiota of a natural Karoo pasture

The survey of a natural Karoo pasture from 1978 to 1982 showed that a wealth and variety o f fungi were present in the semidesert environment. Hyphomycetes and Coelomycetes represented 45.8% and 34.6% respectively of the taxa identified. A total of 135 genera was identified of which Altenaria altemata, Cladosporium spp. and Fusarium spp. of the Hyphomycetes, Phoma spp., Ascochyta spp. and Camarosporium spp. of the Coelomycetes and Leptasphaerulina spp., o f the Ascomycetes represented the most prevalent fungi in this order. This survey has shown conclusively that Pithomyces chartarum, which is associated with photosensitivity diseases o f sheep, can always be recovered from the veld if the correct isolation techniques are employed. A number of new records for South Africa, as well as undescribed species, have been found, highlighting the necessity of correct methods and intensity of approach.


INTRODUCTION
Climatically the Karoo can be defined according to the Holridge System (Price 1975) as semi-arid to arid and warm to cool temperate. Characteristics are erratic, patchy rainfall and occasional unseasonal cold weather when snow may fall in the high-lying areas, even during mid summer ( Figure 1).
The original vegetation of the survey area, which ac cording to Acocks (1979Acocks ( , 1988 was grassland, has largely been replaced by karroid veld, and there is general agree ment that the process of deterioration is continuing with desertification advancing towards the northeast. The flora of the region is rich in species (Acocks 1988) but the habitat is unpredictable with patches of temporary pioneer vegetation (Southwood 1977), comprising species such as Tribulus terrestris L., which become established when the first early summer rains fall.
'Geeldikkop', the hepatogenous photosentivity disease of mainly sheep, was first described by Hutcheon (1886). Theiler (1918) showed that ingestion of T. terrestris, es pecially wilted material, was directly implicated in the aetiology of the disease. He reported the presence of a Colletotrichum sp. on such material and linked it to the disease as a possible cause. According to Watt & Breyer-Brandwijk (1962) the ingestion of T. terrestris causes a condition similar to 'geeldikkop' called 'big head' reported from Colorado and Texas. The plants are high in saponins and thus inherently toxic. In New Zealand (Thornton & Percival 1959;Thornton & Ross 1959) Pithomyces chartarum from ingested grasses proved to be responsible for the development of hepatogenous photosentivity and fa cial eczema, which is very similar to 'geeldikkop* .Very few researchers have been successful in reproducing 'geeldikkop' under field conditions (Van Tonder et al. 1972). Kellerman et al. (1980) were able to show that the combination of P. chartarum and T. terrestris gave histopathological lesions similar to those found during natu ral outbreaks of the disease.
Fungi from litter in the Karoo have received little atten tion. Doidge (1950) reported only a few fungi from the Karoo, mainly collected by MacOwan in the Eastern Cape. On Lycium spp., amongst others, Puccinia lycii Kalchbr. was recorded. No fungi were recorded on Tribulus terrestris L. Pithomyces karoo Marasas & Schumann (1972) was pub lished after a study of litter from the Karoo.
Relatively few surveys of mycobiota have been pub lished from South Africa. Eicker (1973) studied the my-cobiota of Eucalyptus maculata leaf litter. Papendorf & Jooste (1974) described Five species of fungi from wheat field debris after isolation by the dilution plate method. Eicker (1976) studied the mycoflora of Panicum colora tion associated with an outbreak of photosensitivity of sheep for an 11 month period. Bezuidenhout (1977) studied the hyphomycetes (mitotic fungi, Hawksworth et al. 1995) associated with Cenchrus ciliaris L., a fodder grass, over an 11 month period. Van der Merwe et al. (1979) studied the aerospora of an Eragrvstis curvula (Schrad.) Nees pasture in South Africa.
An interim report based on the present survey of Pithomyces chartarum was published stating that a further 315 isolates were tested for sporidesmin production in cul ture of which most did not produce the toxin (Annual Report 1981).
The present survey was initiated to determine the in cidence of P. chartarum in natural Karoo pasture at a time when 'geeldikkop' was likely to occur. The original scope of this study was increased considerably when it became apparent that much valuable information could be gained if a general survey of the mycobiota of the area was done.

MATERIALS AND METHODS
ince. Sampling and monitoring were done over a period of four seasons during which weekly or fortnightly sam ples were collected. The sampling procedure involved tak ing samples from up to seven different plants as well as litter, at three points (1978/79) and later in two camps of a hectare each from 1979 onwards (Table 1).

1978/79 survey
Three sampling points, A, B and C were chosen after completion of a botanical survey of an area where Tribulus terrestris occurred. The nature of the communities at the sampling points varied significantly regarding crown cover, basal cover and density.
Point A was situated in a community with a reasonably high density of perennial Karoo bushes. Therefore, the crown cover was such that wind movement between the individual bushes was possible. The basal cover of T. ter restris was fairly high but decreased with time.
Point B was situated in a very dense community of per ennial Karoo bushes which allowed virtually no wind move ment at soil level. Very few T. terrestris and other pioneer plants, such as Galenia sarcophylla, were present.

Sam pling, m onitoring sites and dates
The survey was conducted at the Grootfontein Agri cultural College Farm, Middelburg, Eastern Cape Prov Point C was situated in an area where only one Lycium cinereum bush of 1.5 m in height was present besides T. terrestris. Virtually no other vegetation was present at this point at the onset of the survey.
Weekly samples of litter were collected, including T. terrestris when present, after rains had fallen during De cember 1978. A total of 34 samples of litter and 20 of T. terrestris plants were studied during this period of seven months.

1979/80 survey
Shifting of plant communities at the points previously chosen necessitated another approach. It was decided to establish two camps (A and B) of one hectare each; the one (A) with a fair cover of T. terrestris, the other (B) without. Five sheep were put into Camp A and the fol lowing plants were sampled: the Karoo bushes Galenia sarcophylla, G. procumbens, Felicia muricata and Lycium cinereum and the grasses Eragrostis lehmanniana and Cynodon incompletus. Other plant materials sampled were unidentified litter and T. terrestris. Initially Camp B con tained very little T. terrestris and, because of well estab lished stands of perennials such as Felicia muricata and Lycium cinereum, was less susceptible to invasion by T. terrestris and other pioneers. The density of the commu nities in Camp B was much higher than in Camp A. and T. terrestris was only found in the comer of the camp adjacent to Camp A. Very few Pentzia spp. and other typi cal Karoo bushes grew in the two camps.
Sampling took place from December 1979 to the end of March 1980 on a fortnightly basis.

1980/81 survey
This survey started in September 1980 and was con tinued through 1981. A total of 52 weekly samples was collected and studied from each of the two camps. This time every plant species named in the 1979/80 survey was, however, sampled and studied individually. Thus four species of bushes, two species of grasses, litter and, when available, T. terrestris were sampled for a full cal endar year.

1981/82 survey
This survey was a continuation of the 1980/81 survey, and continued to the end of March 1982.
2.1.1. Sampling methods employed Samples were taken up to a height of 150 mm which corresponds to the vertical zone grazed by merino sheep. Care was taken to lift litter from the soil surface so as to pick up as few eelworms as possible. The camps were sampled at random to obtain representative samples. If wet, due to rain or dew, the samples were sun-dried before packing into paper bags, every sample from each plant species packed separately, and locality, date and species were noted. Samples were then posted to Pretoria which took approximately 10 days.
The sampling units used were individual leaves, leaf lets and 10 mm lengths of stems and grass blades. The material was sorted and samples from as many different leaves and stems as was possible were taken. Fifty units from each of the samples were planted out directly on potato carrot agar (PCA) (Johnston & Booth 1983) to which 125 mg/1 Albamycin T (Upjohn) had been added prior to autoclaving. Initially some samples from Camps A and B were first washed by shaking in tap water mixed with Teepol (Shell Chemicals) 1:100 in a wrist shaker for 10 minutes to dislodge superficial conidia. The washed material was planted out directly after this treatment. The first five samples collected during the 1980/81 survey were studied this way.
The plates were incubated for a period of seven days at 24°C with intermittent mixed near-UV and daylight fluorescent light from a height of 300 mm on a 12 h/d cycle. The presence of fungi on the material studied was noted and isolations made of P. chartarum and other note worthy fungi. Chemical assays for sporidesmin. the toxin produced by P. chartarum, were done according to the method of Marasas et al. (1972) on a number of the iso lates. Some of these cultures were also used to produce bulk cultures with which to dose sheep.  (Roux 1977). It was later found that the spore trap had to operate too high above the ground to pick up the coni dia released at a much lower level. No spore trap func tioning on a suction principle can operate in a sandy environment at a low level. The use of a spore trapping device was therefore not employed further.

Exposure of Petri dishes
This technique had the dual advantage that it gave the best indication of how many airborne conidia there were, and isolates obtained in this manner were alive and could be used for sporidesmin assays almost right away. How ever, the distance between the sampling site at Grootfontein and Pretoria made this an impracticable method. It was noted that under windy conditions the Petri dishes could be opened for 10 minutes whereas 20 minutes in quiet conditions were needed to give the required results. A larger variety in fungal species was pickcd up in open patches than amongst dense undergrowth. On the lee side of bushes much fewer conidia could be collected. P. char tarum was collected in every Petri dish exposed.

Identification of fungi
The fungi were initially identified at magnifications of 25 x and 50 x using a Zeiss dissecting microscope. Veri fication of identifications was done with a similar make of research microscope. Material was mounted in lactophenol (Johnston & Booth 1983) but from 1980 the coelomycetes (mitotic fungi, Hawksworth et al. 1995) were mounted in ammonium hydroxide with 3.5% erythrosin (Sutton 1980) to facilitate identification based on conidiogenesis. Photomicrographs were obtained using an Olympus microscope camera and Ilford Pan F film.

Meteorological data
Members of the Agricultural Meteorological Division of the Soils and Irrigation Research Institute stationed at Grootfontein recorded and monitored the weather from a casual station in the vicinity. This was equipped with a Stevenson Screen housing a thermohygrograph to record the daily minimum and maximum temperature, an ane mometer, manual and automatic rainfall meters and a grass minimum thermometer.

Veterinary services
Veterinarians stationed at the Regional Diagnostic Laboratory of the Division of Veterinary Services in spected the sheep from time to time for clinical signs of 'geeldikkop'.

Flowering plants sampled
Tribulus terrestris L. (Zygophyllaceae): known as cal trop in the USA, three-cornered jack in Australia and also as Mexican sand-burr (Watt & Breyer-Brandwijk 1962); is notorious for causing disease in sheep and goats; is an annual pioneer plant with a sprawling habit which forms a ground cover; has composite leaves consisting of up to 11 small pinnae covered adaxially with long adpressed unicellular hyaline hairs.
Galenia sarcophylla Fenzl (Aizoaceae): a semisuccu lent, herbaceous ground cover with leaves covered with unicellular and multicellular hairs; is highly palatable to grazing animals and is preferred to T. terrestris; occupies a similar ecological niche as T. terrestris.
Galenia procumbens L.f. (Aizoaceae): a hardy erect shrub about 0.5 m high with small, smooth, simple leaves; is highly palatable to sheep.
Felicia muricata Thunb. (Asteraceae): a multistemmed perennial plant with simple, very small smooth leaves with a sticky surface which serves as an ideal spore trap; is highly palatable to grazing animals.

Lycium cinereum
Thunb. sensu lato (Solanaceae): an erect perennial plant, with woody branches which can reach a height of more than a metre; has simple, smooth leaves and produces small red berrries after flowering in midsummer; in the young stages it is preferentially grazed but is shunned when older and harder, because of its thorny nature; when grazed heavily this species is similar to the smaller Karoo bushes.
Cynodon incompletus Nees (Poaceae): a stoloniferous perennial with a sprawling habit similar to T. terrestris and G. sarcophylla; under adverse conditions the plant is an annual.
Eragrostis lehmanniana Nees var. lehmanniana (Poa ceae): an erect tussock grass which is intensively grazed; usually perennial but it may be annual under adverse con ditions.

Sporidesmin assays
A total of 1005 isolates of P. chartarum were made for toxin production testing. Of these, 437 isolates were se lected and grown on semisynthetic broth (Di Menna et al. 1970) for three weeks under near-UV and daylight fluorescent tubes on a 12 h/d cycle from a height of 300 mm at 20°C. The extraction procedure described by Marasas et al. (1972) was used.

From material directly planted out
All mycobiota identified during this survey are listed in the Appendix. Records of genera and species that were new for South Africa are marked.
The main groups and their incidence in relation to the seasons during the 1980/81 survey are given in Table 2. The total number of genera identified and the percentage representation of classes is given in Table 3. Tables 5, 6 and 7 give complete information regarding the percentage occurrence of the majority of identified fungi on particular substrates for the surveys from 1978 to 1981. Some of the more unusual fungi identified have been illustrated in Figure 2 Table 2 where fungi which occurred continuously can be identi fied as having a peak in a particular season, e.g. summer or winter, as well as on what substrate they occurred. P. chartarum occurred frequently during the first years of   Figure 1.
Average occurrences of the dominant fungi at the vari ous sampling points and areas are presented for the Hyphomycetes (Figure 4), for the Coelomycetes and the genus Leptosphaerulina ( Figure 5), the only ascomycete which occurred continuously for the periods 78/79, 79/80 and 80/81. Sudden fluctuations can be attributed to personal sam pling error when someone other than the regular sampler had collected the samples. (Table 4) P. chartarum does not, under normal circumstances, colonize living leaves in the Karoo and usually occurs as superficial conidia on exposed plant surfaces. Surface ster ilization is therefore not an appropriate technique when looking for this organism. However, the fact that it can occur as an endophyte would add another dimension to its versatility as it is already known as a pathogen of rice (Sutton & Gibson 1977) and a saprophyte.

Sporidesmin assays
A total of 36 isolates or 7.5% of the 1 005 isolates of P. chartarum was positive, and the highest yield was 40 mg/1 sporidesmin. Most isolates, however, gave 10 mg/1 or less sporidesmin under these conditions. The telcomorph Leptosphaerulina chartarum also produced 10 mg/1 sporidesmin under the standard conditions.

Photosensitization
Although Merino sheep were kept in at least one sam pling area at a time, no photosensitization on a clinical level was reported. This is supported by the weather data obtained, which confirmed that no 'danger period' for the outbreak of photosensitization had occurred according to the conditions given by Crawley & Woolford (1965).

Fungi recorded
A significant finding o f this survey was that the Coelomycetes were abundant and diverse and that the number of genera found was nearly equal to that of the Hyphomycetes (Table 5). The 46 genera of identified Coelomycetes (Appendix) and 63 genera of the Hypho mycetes included 24 genera of the Coelomycetes and four genera and 14 species of Hyphomycetes newly recorded for South Africa (see Appendix). Two new records of Ascomycetes were noted, including one new species, Leptosphaerulina chartarum Cec.Roux, which is the teleomorph of Pithomyces chartarum (Roux 1985a).
The total of 63 known genera of Hyphomycetes found in this survey is not as low as it would appear when com pared with other surveys, for example that of Bezuidenhout (1977), which were done on either irrigated lands or under temperate conditions. The fungi in this survey were collected under conditions not usually considered condu cive to the maintenance of an extensive fungal population.
The fungi with consistently high counts were Phoma spp., Alternaria alternata and Cladosporium spp. (Figures  4 & 5). Pugh & Mulder (1971) also encountered Alter naria tenuis, Aureobasidium pullulans, Cladosporium herbarum, Epicoccum nigrum and Phoma typharum as initial colonizers of Typha latifolia L. Populations of Phoma spp. increased over the years which could be due to their being better adapted to the increasingly dry conditions. Ascochyta spp. and Camarosporium spp. increased with time and then levelled off. The incidence of Bipolaris spp. (in cluding related genera such as Drechslera and Exserohilum), Epicoccum nigrum a n d Pithomyces chartarum declined over the study years, although these organisms still occurred consistently. The only Ascomycete which occurred consistently was the genus Leptosphaerulina which also declined eventually (Table 7). It is possible that P. chartarum, which was also present throughout the survey, could have been produced by L. chartarum, which was then counted as P. chartarum rather than as L. char- tarum, when considering the nature of the sporulation straight from the ascospores. The correlation between the incidences of these two fungi, the ana-and teleomorph (mitotic and meiotic, Hawksworth et al. 1995) states, was therefore most significant.
The plant communities studied contain a wealth of fungi, many previously unrecorded. Noteworthy was the occurrence of albino strains of the common species Alternaria altemata, Cladosporium cladosporioides and St a-chybotrys chartarum.
The Hyphomycetes (Table 5) occurred widely and were not as restricted regarding substrate as the other groups en countered. Unusually low incidences were, however, noted for species of Aspergillus, Penicillium and Trichoderma.
The highest incidence of the most prominent genera was noted during autumn and winter (Table 2). This could be explained by the fact that free water in the form of dew and rain was available for longer periods, thus en hancing the growth of fungi. Grass minimum temperatures recorded were substantially lower in winter than in sum mer. Highest rainfall occurred during late summer and autumn, seasons in which the wind tended to subside (Fig  ure 1), thus reducing evaporation.
Nematophagous fungi, such as Dactylella and Candelabrella spp., were found. Large numbers of eelworms were inadvertently picked up with some of the samples and interfered with the counting of the fungi present on the substrate studied.
The entomophagous fungi Beauveria bassiana and Metarhizium cmisopliae were frequently found but only in small numbers. B. bassiana is an important component of a complex of natural enemies of the Karoo caterpillar Loxostege frustalis Zeller (Mohr 1982). During the survey the Middelburg District experienced drought for three suc cessive years. Consequently the ground cover decreased drastically and the unstable sandy soil was disturbed by wind and hoof action. The conidea of B. bassiana. asso ciated with the early subterranean pupal stage of the karoo caterpillar, were therefore set free into the atmosphere in increasing numbers.
The increase in the number of species of Hyphomy cetes from 1980 onwards can also be attributed to the worsening drought conditions which resulted in greater amounts of litter being deposited. The litter became very rich in fungi which would otherwise probably not have been isolated, as the litter fraction represented all the plant material available at the various sampling points and thus included all plant species not sampled separately. It is, therefore, understandable that mycobiota of litter should be much more varied than those of single plant species.
Aspergillus flavus deserves special mention. This toxi genic fungus was very common in animal feeds from all over South Africa examined for mycotoxicological fungi during the entire survey period (Roux 1985b), but it was not recorded in the present survey during the normal rainy season of 1978/79.
In the initial trial run during which material was planted out after washing, P. chartarum was found to be an endophyte. This is even more significant in the light of the subsequent discovery of the teleomorph. Thus P. chartarum, or L. chartarum as it should now be known, can survive unsuitable conditions protected by the leaves of live plants and possibly sporulate when they die. The fact that the conidial stage of L chartarum was found in tissues from all live plants studied is most significant.
Due to the large number of samples and the primary emphasis on Pithomyces chartarum, species of common genera such as Fuscirium, Bipolaris and Leptosphaerulina were not recorded separately. The most common species of Fuscirium was F. moniliforme followed by F. subglutinans. In the Bipolaris group the following species were identified: B. cynodontis, B. halodes, B. hawaiiensis, B.  papendorfii, B. zeicola, Drechslera phlei and Exserohilum rostrata. B. halodes was the most prevalent. Hering (1965) stated that though he had isolated a number of Ascomycetes and Coelomycetes, they failed to grow on the isolation medium. Experience obtained dur ing this study showed that any bacteriostatic agent other than a few drops of lactic acid per Petri dish could com pletely inhibit the growth of some Coelomycetes. This could explain why the numbers of the Coelomycetes re-   1 1.8 0.1 1.1 0.1 1.1 0.1 0.2 0.1 0.1 2.0 2.1 0.1 1.5 1.5 6 1.4 2.2 1.3 1.3 1.2 1.8 4.6 3.9 1.7 0.3 1.1 1.1 0.8  -  1 0.1 0.2 0.3 0.1 0.3 0.1 0.1 0.1 0.2 0.1 0.3 0.1 0.1 0.1 0.1 Dactyella sp.

P hotosensitization
Crawley & Woolford (1965) stipulated a minimum temperature of 12.2°C or more on three consecutive days together with 3.76 mm of rain as a danger period for the development of the facial eczema in sheep. The same con ditions were assumed to be necessary for the development of 'geeldikkop' in local sheep. No such conditions were recorded and no cases of photosensitization on the sam pled pastures were reported.
Another factor which could play a role was the pres ence of saponins in the T. terrestris plants (Watt & Breyer-Brandwijk 1962). Aas & Ulvund (1989) speculated that P. chartarum, especially the sporidesmin present on bog asphodel and saponins, may be involved in the aetiology of alveld (a hepatogenous photosensitivity) in Norway. Since then, Kellerman et al. (1991) have shown that saponins on their own are able to induce hepatogenous pho to sen sitiv ity in som e sheep. The im portance of sporidesmins has, however, not diminished as all sheep in that trial did not react positively. Kellerman et al. (1991) found that fresh T. terrestris, both on its own and with sporidesmin, caused 'geeldikkop' in sheep.

CONCLUSIONS
The survey highlights the wealth and variety of fungi found in this inhospitable environment. The large numbers of genera found is due to the wide range of materials sampled. A peculiarity was that virtually the same number of genera of Coelomycetes and Hyphomycetes was found. Nag Raj (1981) noted that Coelomycetes were more prevalent in dry climates, a fact which has been confirmed here. This phenomenon can be attributed to the adaptation of the fungus in shielding its conidiogenous cells and hya line conidia from the high UV-radiation in the predomi nantly cloudless Karoo region by developing a conidioma. Very few synnematous genera of the Hyphomycetes were recorded. An analogue in the Hyphomycetes is the pro tective mechanism of melanin, because a great proportion of the species present have melanized conidia. This is the first survey in southern Africa in which such a high proportion of fungi identified belonged to the Coelomycetes. The invidual genera could be determined to a great extent using Sutton's keys (1980). Numerous new records for South Africa were registered.
The suitability of litter as a substrate for fungal growth, even under these harsh climatic conditions, was an indi cation of the role fungi play as agents in the breakdown of organic matter. The wide spectrum of fungal genera noted on the litter gave an indication of what was present on substrates not sampled separately.
This survey demonstrated the persistent presence of Pithomyces chartarum on various substrates in the Karoo. This is a very important finding in view of its toxicity. The teleomorph of this fungus, Leptosphaerulina charta rum, was found during this study (Roux 1985a). P. char tarum was recovered from T. terrestris leaves without lesions. This possible endophytic symbiosis of certain strains may indicate its mycotoxicological, opposed to pathogenic (Haware & Sharma 1973) nature and also of the existence of purely saprophytic strains. This survey illustrates the importance of intensive studies of fungal populations.

ACKNOWLEDGEMENTS
The senior author wishes to thank the following for invaluable assistance: the then Director, Karoo Region, Dr P. Roux, who appointed Mr Andries Barnhoorn to assist locally in the taking of samples-the latter did a marvellous job; the late Dr Marius van Tonder, State Vet erinarian, Regional Diagnostic Office, who kept the sheep under surveillance for clinical signs of 'geeldikkop' for five years; the late Mr (Oom Piet) PJ.J. van der Westhuizen, Mycology Unit, PPRI, Pretoria, who painstak ingly and scrupulously prepared the Petri dishes and plated out those thousands and thousands of samples over years; Mrs Marjan Botes (nee De Oude) and Miss Elsie van der Westhuisen, Mycology Unit, PPRI, Pretoria, who assisted with the extraction of sporidesmin; Mr John Pullen, Institute for Climate, Soil and Water, who col lected the meteorological data; the late Mr Jo Minne, who was in the Analytical Chemistry Unit, Toxicology Divi sion, Onderstepoort Veterinary Research Institute, Preto ria, for continual assistance and advice on the extraction of sporidesmin; the late Dr LAP Anderson, for discussions and advice; and lastly my colleagues for undaunting sup port. This paper forms part of a Ph.D degree accepted at the Rand Afrikaans University as submitted by the first author.