Genetic diversity of Claviceps africana on sorghum and Hyparrhenia
Version of Record online: 11 AUG 2005
Volume 54, Issue 6, pages 749–763, December 2005
How to Cite
Pažoutová, S. and Frederickson, D. E. (2005), Genetic diversity of Claviceps africana on sorghum and Hyparrhenia. Plant Pathology, 54: 749–763. doi: 10.1111/j.1365-3059.2005.01255.x
- Issue online: 23 NOV 2005
- Version of Record online: 11 AUG 2005
- Accepted 19 May 2005
- Cerebella andropogonis;
- Claviceps spp.;
RAPD markers were used to survey genetic variability among 140 isolates of Claviceps africana collected from Southern Africa, India, Thailand, Australia and the Americas in 1992–2002. Amplified fragment length polymorphisms were determined for a subset of the isolates. Both markers gave similar results in phenetic analysis of genetic distances between haplotypes of different geographical origin. In the Americas, a single RAPD haplotype was found throughout the various countries. The Eastern lineage consisted of two close haplotypes (one from India, the other from Thailand and Australia). Among five specialized isolates of C. africana from the alternative hosts (Hyparrhenia spp.), three haplotypes were found. Eleven private alleles distinguished the Hyparrhenia population from that on sorghum. rDNA sequences of sorghum and Hyparrhenia isolates differed in three positions. The African sorghum population of C. africana consisted of 10, mostly closely related haplotypes. Low genotypic diversity (HE = 0·0337) and the fact that most of the variation originated from between populations (GST = 0·866) suggested founder effects following recent invasion. In Southern Africa, no significant differentiation was found among six populations. Therefore the data were pooled and tested for prevalence of clonal or sexual reproduction. The presence of the over-represented, widespread RAPD haplotype A; gametic disequilibrium (37% loci detected by exact tests); index of association (IA) significantly >0; and the high proportion of compatible loci (in the clone-corrected and total data sets found to be 94 and 99%, respectively) support the hypothesis of clonality as the predominant means of reproduction.
The oldest reference to ergot disease of sorghum, caused by Claviceps africana, is a report of a famine in 1903–06 in Northern Cameroon, West Africa, attributed to drought and sugary disease of sorghum. Alternatively named Datché (glue) by the local Garoua inhabitants (Beauvilain, 1989), the observed abundance of sticky honeydew points to C. africana as the most likely causal organism. Further intimation of sorghum ergot was from infection of sorghum by Cerebella andropogonis in Kenya in 1923 followed by the deposit of an ergot specimen (IMI 93464) in 1924 (McDonald, 1923, 1925; Wilkinson, 1926): Cerebella is universally recognized as a hyperparasite of, and signal species for, ergot. In 1926 sorghum ergot was collected in Uganda (IMI 14170). Wallace & Wallace (1949) reported ergot infection of sorghum in Tanganyika. In South Africa, a specimen of Claviceps sp. on sorghum was found in Kwa-Zulu Natal in 1936 (specimen no. 28658, Doidge, 1950) and the symptomatology and description of conidia correspond to C. africana (Nowell & McLaren, 1999). A specimen was collected in Nigeria in 1955 (IMI 62801); and Loveless (1964) found conidia of identical shape to C. africana on Sorghum caffrorum in Southern Rhodesia (now Zimbabwe); and Angus collected ergot in Zambia in 1965.
Good pollination of the host is essential to reduce the severity of any ergot infection (Thakur & Williams, 1980; McLaren, 1997). Epiphytotics of ergot disease of sorghum are associated with cold night temperatures (<12°C) 2–3 weeks before flowering and/or cool (c. 20°C), moist weather during the flowering (or bloom) phase of sorghum (McLaren & Wehner, 1992; McLaren, 1997; Wang et al., 2003). The former conditions induce pollen sterility; the latter simultaneously delay pollination and promote proliferation of the pathogen (McLaren & Wehner, 1990, 1992; McLaren & Flett, 1998).
Despite the occasional local outbreak and disease citation, sorghum ergot disease was of no economic importance until the introduction of male sterile (MS) sorghum in the 1960s. Futrell & Webster (1966) recorded the high incidence of ergot in MS breeding lines in Nigeria during 1963–65, and predicted an increase in ergot disease incidence with the more widespread use of MS materials. In the 1980s epiphytotics of sorghum ergot disease across the Southern African region were associated with the expansion of regional F1 hybrid seed production programmes (de Milliano et al., 1991). The means by which C. africana invaded Brazil in 1995 and Australia in 1996 has not been unequivocally established, but the subsequent rapid infection rate and heavy seed losses were facilitated in both cases by cool, humid weather at flowering (Bandyopadhyay et al., 1998).
The sudden, rapid, worldwide spread of C. africana generated interest in the infraspecific variation of this economically important sorghum pathogen, and it was subsequently analysed in number of studies. Pažoutováet al. (2000a) compared American, African, Indian and Australian isolates using random amplified polymorphic DNA (RAPD) analyses; and Tooley et al. (2000, 2002) analysed the variability of predominantly American isolates by amplified fragment length polymorphisms (AFLP) and random amplified microsatellites (RAM). Komolong et al. (2002) focused on Australian isolates and analysed radiolabelled DNA amplification fingerprints (RAF). Although the methods differed in the amount of infraspecific variation detected, all these papers concluded that the American invasion clone and its descendants were more closely related to recent African clones than the Asian–Australian populations.
In the present study, the variation of C. africana in Southern Africa and its changes over the period 1992–2002 were assessed using RAPD analysis, with AFLP analysis on a smaller subset of isolates. RAPD data were used to estimate gene diversity and population differentiation. Multilocus analysis of C. africana populations from Southern Africa was performed to determine whether clonal or sexual propagation predominates in this species. Claviceps africana demonstrates an enormous propensity to produce (asexual) windborne secondary conidia (Frederickson et al., 1993) that travel moderate distances, but the perfect stage of the pathogen has proven difficult to generate (Frederickson et al., 1991).
Among non-Sorghum grasses, Panicum maximum (Urochloa maxima) (Futrell & Webster, 1966) and pearl millet (Frederickson & Mantle, 1996) were reported as susceptible to C. africana. In contrast, Reed et al. (2002) did not observe disease symptoms after artificial inoculation of several millets (finger millet, pearl millet, proso millet, foxtail millet) and weed grasses (Andropogon scoparius, Sorghastrum nutans, Panicum virgatum and Elymus canadensis) in a glasshouse experiment. Muthusubramanian et al. (2005) successfully inoculated Sorghum arundinaceum, Sorghum halepense, Sorghum versicolor, Sorghum virgatum and Pennisetum glaucum with conidial suspensions of C. africana and also C. sorghi. The influence of host plant on the shape and size of conidia produced by sphaceliae of both species was observed.
Grasses in the neighbourhood of sorghum fields were therefore surveyed for ergot infection, indicative of a possible role as an alternative host of C. africana and, potentially, a source of further genetic variability of the pathogen.
Additional data concerning C. africana distribution before the 1960s, particularly occurrences before 1964, were obtained from herbarium specimens of C. andropogonis (a signal species for ergot infections) on sorghum from Brazil, Liberia, Sudan, Mozambique and the Philippines (US National Fungus Collections, BPI) and the morphology of Claviceps conidia under Cerebella sporodochia was observed.
Materials and methods
Claviceps africana isolates
Over 200 isolates of C. africana were studied from three African countries (South Africa, Zimbabwe and Zambia), Australia, India, Thailand, Brazil, USA (Texas) and Mexico (Table 1). Selected representatives of the Australasian and American populations used in previous studies (Pažoutováet al., 2000a; Komolong et al., 2002; Tooley et al., 2002) were included for continuity and comparison. Ergotized sorghum specimens were collected during the period 1999–2002 in Southern Africa. Sample sites were chosen to represent locations where natural C. africana disease epiphytotics had occurred previously in farm or research station sorghum. Sampling of African isolates at 10–20 m intervals was in a diagonal transect pattern (zig-zag) across each site to include, where possible, different sorghum varieties. Where ergot severity permitted, sampling was also from different infected spikelets on the same panicle. Additionally, three isolates were obtained from frozen honeydew samples that had been collected in Matopos (Zimbabwe) in 1992. Isolates were made by plating honeydew from each sorghum sample onto T2 sucrose–asparagine agar as previously described (Pažoutováet al., 2002a).
|Haplotype||Isolatesa||Tooley et al. (2000, 2002)||Komolong et al. (2002)||Origind||Year|
|A||92-4, 5, 6||MRSc, Matopos, nr Bulawayo, ZW||1992|
|A||AWH1||C1||Cla 120, 121||Harare, ZW||1999|
|A||AWH1||Africa 1||Cla 56, 58||Cla 58||Potchefstroom, NW, SA||1998|
|A||AWH3||Afr 2·2||Hazyview, Mpumalanga, SA||1999|
|A||3 isolates/1 sample||Balfour, Gauteng, SA||2001|
|A||2 isolates/2 samples||Deel Kraal, Potchefstroom, NW, SA||2001|
|A||1, 3, 4, 5||ARC, Potchefstroom, NW, SA||2001|
|A||2, 4, 5, 6||Pannar Seed Co., Klerksdorp, NW, SA||2001|
|A||1·2, 2·2, 3, 6·1||ARC, Bethlehem, Free State, SA||2001|
|A||6 isolates/3 samples||Sasolburg-Heilbron, Free State, SA||2001|
|A||15 isolates/5 samples||Viljoenskroon, Free State, SA||2001|
|A||5, 6, 7||Cedara, Kwazulu Natal, SA||2001|
|A||6 isolates/6 samples||MRS, Matopos, nr Bulawayo, ZW||2000|
|A||2 isolates/1 sample||Lucydale Farm, Matopos, nr Bulawayo, ZW||2000|
|L||10 isolates/9 samples||MRS, Matopos, nr Bulawayo, ZW||2000|
|L||6 isolates/4 samples||Lucydale Farm, Matopos, nr Bulawayo, ZW||2000|
|L||10 isolates/7 samples||Aisleby Farm, Bulawayo, ZW||2000|
|L||6 isolates/1 sample||MRS, Matopos, nr Bulawayo, ZW||2001|
|L||28 isolates/19 samples||Lucydale Farm, Matopos, nr Bulawayo, ZW||2001/2|
|B||1·1, 6·2||ARC, Bethlehem, Free State, SA||2001|
|Ca||3||Cedara, Kwazulu Natal, SA||2001|
|Cb||8||Cedara, Kwazulu Natal, SA||2001|
|P||3||Pannar Seed Co., Klerksdorp, NW, SA||2001|
|P||2||ARC, Potchefstroom, NW, SA||2001|
|F||F||F1||Cla 122||Bulawayo, ZW||1999|
|S||S||E3||Cla 128||Panmure Farm, Shamva, ZW||1999|
|D||D||D1, 2||Cla 89||Golden Valley Res. Stn, Chisamba, Zambia||1999|
|W||AWH2||Bol 1, 2||Bolivia||1996|
|W||E 267·1–3||Kelly Green, Texas, USA||1997|
|W||AWH4||KG||KG||Kelly Green, Texas, USA||1997|
|W||A 357·1, 357·2||TAMU Research Farm, College Station, Texas||1997|
|W||A 623·1, 623·2||College Station, Texas||1997|
|W||TAM||TAMU Research Farm, College Station, Texas||1997|
|W||AWH5||MX 9·2||Cla 66||Celaya, Guanajuato, Mexico||1998|
|W||AWH2||MX 2·1||Celaya, Guanajuato, Mexico||1998|
|W||AWH5||PrCI, Pr8||Cla 32, Cla 34||Yela Farms, Guyanilla, Puerto Rico||1998|
|W||AWH2||Afr 2·1, 2·3, 3·3||Hazyview, Mpumalanga, SA||1999|
|W||AWH5||Braz 3, 4||Passo Fundo, Brazil||1999|
|W||3 isolates/1 sample||St Catarina, Brazil||2002|
|E||EH2||Indore||Indore, Madhya Pradesh, India||1998|
|E||EH2||Akola||Cls 2||Akola, Maharashtra, India||1998|
|E||Dharwar||Dharwar, Karnataka, India||1998|
|E||EH2||NAP3||Nalgonda, Andhra Pradesh, India||2000|
|E||EH1||NAP10||Adilabad, Andhra Pradesh, India||2000|
|E||EH2||AP14||AP16||Mahbubnagar Andhra Pradesh, India||2000|
|E||APAU1||APAU 1||Ranga Reddy, Andhra Pradesh, India||2000|
|E||EH5||KA61b||Bidar, Karnataka, India||2000|
|E||EH4||KA61c||KA61b||Bidar, Karnataka, India||2000|
|E||EH1||SK3, SK5||SK5||Mandya, Karnataka, India||2000|
|E||EH1||SK7||SK7||Hassan, Karnataka, India||2000|
|E||EH1||SK13||Chitradurga, Karnataka, India||2000|
|E||EH4||MH70||Cla IN29||MH70||Akola, Maharashtra, India||2000|
|E||EH4||MH71||MH71, MH72||Nagpur, Maharashtra, India||2000|
|E||EH2||MH76||Nanded, Maharashtra, India||2000|
|E||EH2||MH79||Parbhani, Maharashtra, India||2000|
|E||EH3||NI1||Rampur, Uttar Pradesh, India||2000|
|E||EH1||NI2||Jhansi, Uttar Pradesh, India||2000|
|E||EH3||NI3||NI3||Jhansi, Uttar Pradesh, India||2000|
|E||EH3||NI6||NI6||Nainital, Uttar Pradesh, India||2000|
|E||EH3||NI9||Rampur, Uttar Pradesh, India||2000|
|E||EH3||NI10||Moradabad, Uttar Pradesh, India||2000|
|E||EH3||NI12||NI12||Udaipur, Rajasthan, India||2000|
|E||G11||Bidar, Karnataka, India||2001|
|Ea||EH2||(3 isolates 1 sample)||Maung Distr., Lap.Buri Prov., Thailand||1999|
|Ea||EH1||AU||Mutchilba, Queensland, Australia||1996|
|Ea||T10763||Cla 1, 2||T10763||Norwin, Queensland, Australia||1996|
|Ea||EH1||T10764||Cla 1, 2||Norwin, Queensland, Australia||1996|
|Ea||EH2||T10765||T10765||Mutdapilly, Queensland, Australia||1996|
|Ea||T10777||Kingaroy, Queensland, Australia||1997|
|Hyp a||Hyp||HypZW00 1, 2, 3||MRS, Matopos, nr Bulawayo, ZW||2000|
|Hyp b||HypZW01||MRS, Matopos, nr Bulawayo, ZW||2001|
|Hyp c||HypSA01 1, 2||Cedara, Kwazulu Natal, SA||2001|
Search for alternative hosts of Claviceps africana
Grasses in the vicinity of sorghum fields in Zimbabwe and South Africa, including Hyparrhenia rufa, Hyparrhenia tamba and Loudetia flavida, were scrutinized for signs of Claviceps infection. The shape and size (morphology) of honeydew conidia were compared to that of C. africana and other available known Claviceps spp. from the same host genus. An additional comparison of RAPD and ITS-5·8S rDNA sequence data was facilitated when the conidia in honeydew, or the sclerotia, were viable. Isolates were obtained from honeydew (Pažoutováet al., 2002a) or following the surface sterilization of sclerotia by 2% sodium hypochlorite (diluted Clorox) for 4 min, two rinses with sterile distilled water and plating on T2 agar.
Spore mounts of macro (primary) conidia from Claviceps infections on sorghum and other grasses were made by diluting honeydew droplets in cotton blue/lactophenol and mixing with polyvinylalcohol. The length and width of 50–200 conidia from infections on sorghum, H. rufa and H. tamba were measured using an Olympus BX51 microscope equipped with a digital camera (CAMEDIA) and image-processing software (quickphoto micro 2·0). Five samples (panicles) per host/location were observed. Sorghum and Hyparrhenia means were compared using the Tukey–Kramer test as implemented in kyplot ver. 2·0 beta 15 (1997–2000, Koichi Yoshioka). Five herbarium specimens of Sorghum spp., overgrown with C. andropogonis, were obtained courtesy of US National Fungus Collections (BPI), spore mounts were made and conidial lengths and widths recorded.
DNA preparation, RAPD and rDNA sequencing methods used in this study are described in detail elsewhere (Pažoutováet al., 2000b, 2002a). For RAPD analysis, primers OPA01 (CAGGCCCTTC), OPA09 (GGGTAACGCC), 8F (GCTCTGAGATTGTTCCGGCT), 30F (GAGGACGATTCATCAACC), 5R (TTTGTCCGGCTCAGAAAC), 10R (GGCCAGTGTGAATATGC) and ITP5 (CCGGCTTGTATTGG) were used. AFLPs were generated as described by Zeller et al. (2000) and Vos et al. (1995). Genomic DNA (100 ng) was digested by EcoRI and MseI and ligated to adapters, preamplified using primers EcoRI-core (CTCGTAGACTGCGTACCAATTC) and MseI-core (GACGATGAGTCCTGAGTAA) and amplified with final amplification primer pairs EcoRI + AG/MseI + C (Tooley et al., 2002), EcoRI + TT/MseI + AC, and EcoRI + GG/MseI + CT. EcoRI primers were end-labelled with γ33P and fragments were separated in 6% polyacrylamide gel (Long Ranger FMC, USA). Polymorphic AFLP and RAPD fragments were scored as binary characters for each isolate, whereas monomorphic fragments across all isolates were not included. Thirty-five RAPD bands and 39 AFLP bands were scored. These were phylogenetically informative in that they were shared by two or more isolates, or lacking from two or more other isolates. A distance matrix of isolates was calculated for both fingerprinting methods according to Nei & Li (1979) with 500× bootstrapping, and a dendrogram was constructed using the unweighted pair group method with arithmetic mean (UPGMA) as implemented in treecon 1·3b (van de Peer & de Wachter, 1997).
Diversity statistics for each group of populations or single population included: number of haplotypes; number of polymorphic loci and their percentage using a 95% criterion; and number of private alleles. The genotypic diversity (equivalent to expected heterozygosity in diploids) was calculated as HE = n/n − 1(1 − Σ ), where pi is the frequency of the ith genotype and n is the total number of isolates sampled (Nei, 1987). This ratio approaches 0 if all individuals sampled have identical genotypes, and equals 1 if each individual has a unique genotype. Shannon's diversity index (Lewontin, 1972) was defined as I = –Σ pi ln pi (i = 1,2), where pi was the frequency of the presence or absence of a given RAPD band. In addition, GST (a proportion of total variation that is distributed among populations, for haploids equal to FST) (Nei, 1987) was calculated. Diversity calculations were performed using popgene32 (Yeh et al., 2000). Genetic distances were calculated using Nei's unbiased genetic distance (Nei, 1978) and the topology was inferred by the UPGMA with bootstrap. The relationship between the genetic distance matrix [Wright's (1978) modification of Roger's distance] and log geographic distance matrix was evaluated with a standardized Mantel statistic (Sokal & Rohlf, 1995). All calculations were performed using tfpga (Tools for Population Genetic Analyses; Miller, 1997). Population differentiation was assessed using Weir & Cockerham's (1984) population differentiation theta statistic performed on clone-corrected and uncorrected data sets with 10 000 randomizations (multilocus 1·2, Agapow & Burt, 2000) and by exact tests (Raymond & Rousset, 1995a) as implemented in tfpga (Miller, 1997). Gene flow was estimated as Nem, the product of effective population size and immigration rate, and is referred to as the effective number of migrants per generation. Nem was calculated using private alleles (Barton & Slatkin, 1986) as implemented by genepop web version 3·4 (Raymond & Rousset, 1995b). Random association among loci as a sign of recombination (Maynard Smith, 1999) was tested by several methods on clone-corrected and total data sets. In a clone-corrected data set, each haplotype was included only once per population. Variance, VD, was calculated from a distribution of mismatch values derived from differing pairs of loci. This was compared with the variance expected in the case of linkage equilibrium. The null hypothesis, VD = VE, was tested by a Monte Carlo simulation and a parametric test in lian 3·1 (Haubold & Hudson, 2000). Index of association, IA (IA = VD/VE − 1) was calculated from the variances obtained using lian 3·1. IA is a function of the rate of recombination and its value is 0 in the presence of random mating (Brown et al., 1980). The proportion of compatible pairs of loci was calculated using multilocus 1·2 (Agapow & Burt, 2000). Two loci are compatible if all the observed genotypes are explainable by mutation only; no homoplasy or recombination is inferred (Estabrook & Landrum, 1975). In the gametic (linkage) disequilibrium test, the null hypothesis states that genotypes at one locus are independent of genotypes at the other locus. The program arlequin (Schneider et al., 2000) runs a Fisher's exact probability test on contingency tables using a Markov chain; it was performed under the default parameters (Raymond & Rousset, 1995a).
More than 200 isolates worldwide were studied, mostly from sorghum. Five isolates of C. africana were found on Hyparrhenia spp. (see below). A total of 15 haplotypes (multilocus genotypes) were found, 13 of them in Africa (Fig. 1). Australian isolates all belonged to the same haplotype, Ea, differing from the common Indian haplotype, E, by a single band acquisition, detectable with the primer ITP5. The same haplotype was also found in Thailand, but not in India.
All the American isolates belonged to a single haplotype, W, which was also detected in a honeydew drop from a sorghum head originating from Hazyview (South Africa). Another Hazyview isolate from the same sorghum panicle belonged to the most widespread African haplotype, A. All isolates of the American group shared three distinctive bands that were absent from any of the remaining African haplotypes, supporting the idea of their common origin from a single invasion clone that entered Brazil in 1995.
The most widespread African haplotype, A, was first detected at Matopos Research Station (Bulawayo, Zimbabwe) in 1992. It was later reisolated in Zimbabwe in the Bulawayo region (including Matopos Research Station, Aisleby Farm and Lucydale Farm), near Harare, as well as from various locations in South Africa. However, it has now been replaced in the Bulawayo region by a derived haplotype, L, differing by an additional band with the primer OPA 09. The new haplotype was first detected in the year 2000, when 26 isolates belonging to haplotype L, and only eight to the original haplotype A, were found. In the year 2001 only haplotype L was found.
The African haplotypes D, S and F were each found only once, and F (from the Bulawayo region) was never reisolated despite extensive collections in that area.
The most distant and highly supported clade consisted of haplotypes Ha, Hb and Hc found on an alternative host genus, Hyparrhenia.
A subset of C. africana isolates was analysed by AFLP (Figs 1 and 2). African and American haplotypes, AWH1–5, were closely related and clustered in a highly supported clade. African isolates belonging to the RAPD haplotype A exhibited two AFLP haplotypes, AWH1 and AWH3. In the African isolate from Hazyview with RAPD haplotype W, haplotype AWH2 was found (Table 1). In the American isolates, haplotypes AWH2, AWH4 and AWH5 were observed.
The lineage of haplotypes from Australia and Asia was clearly separated from the American and related African lineages, and consisted of six haplotypes. Haplotypes EH1 and EH2 were found in Thailand and Australia. In India, haplotypes EH2–EH6 occurred. Contrary to the RAPD, no clear-cut haplotype differences between Indian and Australian isolates were found. However, more variation was found in this lineage than among the American isolates, especially with the primer pair EcoRI + TT/MseI + AC.
Haplotypes S and F (Zimbabwe) shared a clade, distant from all other groups. However, in contrast to the RAPD analysis, haplotype D (Zambia) was related more to these two haplotypes than to the American one. The most distant clade contained the haplotype from H. rufa.
Population diversity of Claviceps africana
Isolates from African locations were grouped into six populations according to the sorghum-growing region from which they originated – Matopos, Hazyview, Cedara, Bethlehem, Potchefstroom (including locations at Potchefstroom, Deel Kraal and Viljoenskroon), and Balfour (including samples from the area between Sasolburg and Heilbron) (Table 2). All the populations shared a multilocus haplotype, A, that was also the only haplotype found in the Balfour area. Private multilocus haplotypes were found in five populations, whereas private alleles occurred in three of them.
|Matopos||52||3||9||4||0·019 (0·017)||0·0372 (0·0964)|
|Hazyview||2||2||6||4||0·171 (0·185)||0·1188 (0·2650)|
|Cedara||5||3||4||1||0·046 (0·038)||0·0572 (0·1615)|
|Bethlehem||6||2||1||0||0·015 (0·017)||0·0182 (0·1076)|
|Potchefstroom||27||2||1||0||0·004 (0·007)||0·0075 (0·0446)|
|Pooled populations||101||8||15||0·026 (0·020)||0·0372 (0·0964)|
The differences among the six populations for the average genotypic diversity were not significant due to their high standard errors. Again, the values for gene diversity, H, were very low. Genetic distances for these populations were small (Table 3). Nei's unbiased genetic distance among all population pairs ranged from 0·0001 between Potchefstroom and Sasolburg to 0·047 between Matopos and Hazyview, again indicating no strong differentiation. No significant overall population differentiation was found among the six populations based on the theta statistic (θ = −0·2156, P = 0·952). Similarly, exact tests for population differentiation (Raymond & Rousset, 1995a) revealed no significant differences (P in the range 0·639–1·0). The correlation between genetic and geographic distance was positive (0·5880) but not significant (Mantel test, P = 0·084).
Worldwide, isolates of C. africana differed in 35 polymorphic loci which were used for the calculations in Table 4. The most diverse population of C. africana from sorghum is in Africa, which correlates with an African centre of origin. Six private alleles typical for the continent and 10 haplotypes were found. In the Americas a single haplotype, W, was found throughout the various countries sampled. The Eastern lineage consisted of two closely related haplotypes, one from India, the other found in Thailand and Australia; none of them was recovered from among the recent African collections. The low gene diversity of the non-African lineages is explained by the founder effect of the invasion clones.
Although only five isolates of C. africana from Hyparrhenia spp. were collected, there were three haplotypes found. Eleven private alleles distinguished the Hyparrhenia population from that on sorghum.
The main source of variation (measured by GST) in the populations from Sorghum or Hyparrhenia with more haplotypes was between collection sites. Based on Nei's unbiased distances between populations (Fig. 3), the population distances on the UPGMA dendrogram corresponded approximately to the supposed time interval since their separation. Genotypic diversity, HE, was low, reflecting the small number of haplotypes encountered overall.
The effective number of migrants was very low between any of the populations compared (Table 5), indicating that after the initial invasion and founding epiphytotic, almost no further contact occurred.
|Gene flow||Nema||Mean frequency of private alleles|
|Among six sorghum populations||0·115||0·305|
|Between Africa and Americas||0·0367||0·430|
|Between Africa and Australasia||0·029||0·503|
|Between African sorghum and Hyparrhenia populations||0·019||0·672|
|Between Americas and Australasia||0·007||0·977|
Potential for clonal reproduction
The presence of a particular multilocus haplotype in great abundance in distant localities, or in samples taken years apart, is the most significant evidence for clonal reproduction (Tibayrenc et al., 1991). Isolates with haplotype A were collected from Matopos (Zimbabwe) in 1992 and 2000, and throughout South Africa in 1999–2001. Thus, this clone appears to be widespread in Southern Africa over several years. Among the 101 isolates in this data set, only eight haplotypes (or multilocus genotypes) were found. Haplotype A has been found 53 times, followed by L (40 times), prevailing in Matopos during 2001–02. All the other haplotypes were found only once or twice at the given location. This was reflected by overall low values of genetic diversity, HE. Six populations from Southern Africa were pooled (as their differentiation was not significant) and used for clonality testing, although the sampling was not designed primarily for this purpose.
Out of the six populations, gametic (linkage) disequilibrium was found only in Matopos. In the remaining five it was either not detected, or impossible to calculate due to the presence of only one (or no) polymorphic locus. However, in the data set of pooled populations, 39 significant linkage disequilibria (37% of pairings) were found with Fisher's exact test.
The observed variance, VD (Table 6), was calculated from a distribution of mismatch values derived from differing pairs of loci. This was compared with the variance expected in the case of linkage equilibrium. The null hypothesis, VD = VE, was tested using a Monte Carlo simulation (Haubold & Hudson, 2000). In both original and clone-corrected data sets, the VD value was higher than VE and was well outside the 95% confidence limit. The IA was significantly higher than zero in both original and clone-corrected data sets.
|VDa||Var (VD)||VEb||Lc||IAd||Compatible loci (%)|
The proportion of compatible pairs of loci in the clone-corrected and total data sets was found to be 94 and 99% of compatible loci, respectively. This confirms a lack of recombination and suggests mutation as the main source of genotype variation.
The presence of over-represented, widespread genotype A, linkage disequilibrium (exact test and IA), and the high proportion of compatible loci support the hypothesis of clonality as the predominant means of reproduction in C. africana.
Screening for alternative hosts of Claviceps africana
Several ergotized grass species were collected adjacent to sorghum fields (Table 7). The morphology of sphacelial macroconidia (Loveless, 1964) with data from DNA analyses revealed in most cases that the causal agent was not C. africana.
|Host||Claviceps species||Locationb||Accession no.c|
|Melinis repens||C. rhynchelytri||Matopos (ZW)||AJ605998|
|Cynodon dactylon||C. cynodontis||Matopos (ZW)||AJ557074|
|Heteropogon contortus||Unknown, ellipsoid macroconidia||Matopos (ZW)||NDd|
|H. contortus||C. pusilla||Matopos (ZW)||AJ537577|
|Bothriochloa insculpta||C. pusilla||Matopos (ZW)||ND|
|Hyparrhenia sp.||C. pusilla||Arsi Negele, Ethiopia||ND|
|Hyparrhenia spp.||Conidial groupa 11||Matopos (ZW)||AJ537576|
|Klerksdoerp, Deel Kraal (SA)|
|Loudetia flavida||Conidial groupa 13||Matopos (ZW)||AJ605997|
|Urochloa maxima||C. maximensis||Matopos (ZW)||AJ133396|
|Hyparrhenia rufa||C. africana||Matopos (ZW)||AJ605994|
|Hyparrhenia tamba||C. africana||Cedara (SA)||ND|
Despite the presence of Claviceps pusilla (Fig. 4a) near the sorghum fields and its known wide host spectrum among andropogonoid grasses (Langdon, 1954), this species has never been encountered on sorghum.
On several unidentified species of Hyparrhenia, a species of Claviceps with characteristically oblong, truncated conidia (Fig. 4b) was often observed. This pathogen was similarly isolated from various Hyparrhenia species, as well as from Andropogon gayanus, by Loveless (1964, 1985; his group 11). As Loveless never observed or produced the teleomorph, he could not identify the species. The rDNA sequence of the species was 97·6% identical with that of C. pusilla, making it a sister species to this taxon.
The macroconidia from L. flavida (Matopos, Zimbabwe) (Fig. 4c) strongly resembled those of C. africana in shape (conidial group 10, according to Loveless, 1964), but were smaller (10·3 × 4·4 µm). The sequence data of the rDNA proved that this species was not C. africana. It appears to be identical to Loveless's conidial group 13 (Loveless, 1964).
Claviceps africana var. hyparrheniae
Macroconidia from ergotized H. rufa (Matopos, Zimbabwe) and H. tamba (Cedara, South Africa) (Fig. 4d) had the distinctive shape of C. africana macroconidia from Sorghum bicolor, and the length and width measurements (Table 8) were at the upper end of the range previously recorded for C. africana (Loveless, 1964; Frederickson et al., 1991). Pear-shaped secondary conidia, characteristic of C. africana, were produced in nature on the surface of honeydew in both hosts, and microconidia were seen in addition to macroconidia. Macroconidia from H. rufa and H. tamba (mean of three specimens, length 17·02 ± 0·59; width 7·5 ± 0·55 µm) were significantly longer and wider than those from S. bicolor (mean of 27 specimens, length 12·7 ± 1·02; width 6·4 ± 0·47 µm) (Tukey–Kramer test: P < 0·001 in both cases).
|Locationa||Year||Lengthd (µm)||Widthd (µm)|
|Sorghum||1999||11·4 ± 1·4||5·7 ± 0·6|
|Harare, ZW||1999||11·4 ± 1·4||5·7 ± 0·6|
|Hazyview, Mpumalanga, SA||1999||13·0 ± 1·6||6·0 ± 0·5|
|Bulawayo, ZW||1999||13·4 ± 1·9||5·8 ± 0·8|
|Panmure Farm, Shamva, ZW||1999||13·9 ± 1·5||7·4 ± 0·7|
|Golden Valley Research Station, Chisamba, Zambia||1999||13·9 ± 1·8||6·1 ± 0·7|
|MRS, Matopos, nr Bulawayo, ZW||2000||11·2 ± 1·4||5·8 ± 0·6|
|11·7 ± 1·3||6·2 ± 0·6|
|Lucydale Farm, Matopos, nr Bulawayo, ZW||2000||12·5 ± 1·1||6·3 ± 0·6|
|Aisleby Farm, Bulawayo, ZW||2000||13·0 ± 1·2||6·7 ± 0·6|
|13·1 ± 1·5||7·0 ± 0·6|
|Cedara, Kwazulu Natal, SA||2001||14·3 ± 1·4||6·9 ± 0·7|
|Balfour, Gauteng, SA||2001||13·8 ± 1·6||6·2 ± 0·5|
|Deel Kraal, Potchefstroom, NW, SA||2001||13·2 ± 1·5||6·1 ± 0·5|
|ARC, Potchefstroom, NW, SA||2001||13·3 ± 1·6||6·0 ± 0·5|
|Pannar Seed Co., Klerksdorp, NW, SA||2001||13·8 ± 1·5||6·2 ± 0·5|
|ARC, Bethlehem, Free State, SA||2001||13·3 ± 1·7||6·7 ± 0·6|
|Sasolburg-Heilbron, Free State, SA||2001||12·6 ± 1·7||6·6 ± 0·7|
|Viljoenskroon, Free State, SA||2001||13·8 ± 1·6||6·7 ± 0·6|
|Passo Fundo, Brazil||1999||12·0 ± 1·9||5·7 ± 0·6|
|Passo Fundo, Brazil (S. halepense)||1999||11·1 ± 1·2||6·4 ± 1·7|
|Corpus Christi, Texas||2002||12·1 ± 1·6||6·3 ± 1·9|
|Corpus Christi, Texas||2004||13·3 ± 1·8||6·5 ± 0·6|
|Miyakonojo, Miyazaki, Japan||1991||11·6 ± 1·2||6·5 ± 0·4|
|Miyakonojo, Miyazaki, Japan||1993||11·7 ± 1·3||6·6 ± 0·5|
|Bundaberg, Queensland, Australia (BRIP 23521)||1996||14·0 ± 1·3||7·5 ± 0·6|
|Maung District, Lap Buri Province, Thailand||1999||11·6 ± 1·9||6·4 ± 0·7|
|Bidar, Karnataka, India||2001||11·5 ± 1·6||6·1 ± 0·7|
|Bidar, Karnataka, India||2001||11·4 ± 1·6||5·8 ± 0·6|
|Hyparrhenia||2000||17·3 ± 1·4||7·0 ± 0·4|
|MRS, Matopos, nr Bulawayo, ZWb||2000||17·3 ± 1·4||7·0 ± 0·4|
|MRS, Matopos, nr Bulawayo, ZWb||2001||16·3 ± 1·3||7·5 ± 0·7|
|Cedara, Kwazulu Natal, SAc||2001||17·4 ± 1·4||8·1 ± 0·7|
AFLP patterns with primers EcoR1 + TT/Mse1 + AC (Fig. 5) and RAPD patterns with primer 10R and 8F confirmed isolates from both Hyparrhenia species as C. africana. However, critical differences from representatives of the sorghum population were observed in RAPD and AFLP patterns with all other primers. The rDNA sequences of sorghum isolates from Bolivia, India and Australia (AJ011590, AJ011784 and AJ011784, respectively) were identical (Pažoutováet al., 2000a; Komolong et al., 2002). In the rDNA sequence AJ605994 of the Hyparrhenia isolate, three differences were found in the ITS2 region: two C–T transitions and an indel (gap).
In addition to DNA data and conidial size differences, morphological differences were observed between sphacelia/sclerotia (Fig. 6a): those from sorghum were globose, white internally, with a thin, orange-brown to red-brown rind (Frederickson et al., 1991), but those from the Hyparrhenia species were elongate and cylindrical. This difference may be explained by sphacelia/sclerotium assuming the shape of the seed being replaced. All sclerotia from Hyparrhenia looked fully mature, with a differentiated black rind and protruding sphacelial cap (Fig. 6b), unlike those from sorghum.
Hyparrhenia isolates retained conidiation ability in culture on T2 agar medium, whereas sorghum isolates regularly lost this ability during growth of the first colony in the course of the isolation procedure. Not one isolate with characteristics of the Hyparrhenia pathogen was found on neighbouring sorghum plants, either in Matopos (Zimbabwe) or in Cedara (South Africa), and vice versa.
In Australia, Langdon (1952) observed Claviceps inconspicua conidia c. 15–20 × 5–10 µm, with straight or slightly rounded sides and with rounded ends. He found very similar conidia on a specimen of ergotized Hyparrhenia sp. from Nyassaland (today Malawi) (IMI35739 from 1949) which, in the light of recent observations, might well be those of Hyparrhenia-specialized C. africana.
All this evidence suggests a specialized population or emerging subspecies that was assigned the name Claviceps africana var. hyparrheniae.
The fungus Cerebella is regarded as a signal species for ergot (Langdon, 1955). Thus the Brazilian record of Cerebella on Sorghum in 1932 suggested the presence of sorghum ergot well before the first official record (Reis et al., 1996) and ran counter to global opinion on its absence in the Americas until 1996. However, it was found that the sample had been misidentified and that it contained only Epicoccum nigrum, macroscopically mimicking the Cerebella sporodochium. No ergot infection was found under the sporodochium, confirming the first record of C. africana on sorghum in Brazil as 1996 (Reis et al., 1996).
In all other herbarium specimens (Table 9), ergot infection under Cerebella sporodochia was found. The 1948 Liberian sample contained a few immature C. africana sclerotia of typical globose morphology, and had been obtained from an import of sorghum grain (S. bicolor) by the USA phytosanitary control. However, no ergot sample of the same origin was deposited simultaneously, suggesting that ergot infection went unnoticed.
|US0455076||Minas Gerais, Brazil||S. vulgare||Epicoccum nigrum||1932||–||–|
|US0455138||Belijirai, East Province, Liberia||Sorghum||Claviceps africana||1948||14·0 ± 1·7||4·9 ± 0·6|
|US0455139||Doleil Hill, Sudan||S. lanceolatum||C. africana||1930||13·4 ± 1·7||5·8 ± 1·0|
|US0455011||Mozambique||S. vulgarea||C. africana||nd||13·0 ± 1·7||5·2 ± 0·6|
|US0455012||Los Baños, Philippines||S. vulgare||Claviceps sp.||nd||14·6 ± 2·0||5·9 ± 1·1|
Macroconidia observed in the African specimens corresponded to the description of C. africana (Fig. 7a). The Mozambiquan sample also contained many rounded microconidia (2·5–5 µm diameter). The shape of Sudanese conidia was the most similar to recent C. africana spores. Sclerotia and sphacelia of African samples were globose, resembling the shape of sorghum seed (Fig. 7c).
The specimen from the Philippines also contained Cerebella-free florets. It was not dated but, from details in the accompanying notes, it originated well before 1950. The conidia had an elongated to slightly clavate, or almond-like, shape (Fig. 7b); sclerotia were elongated, unlike those from the Mozambiquean sample (Fig. 7c) on the same sorghum species. In contrast to C. africana infections, the glumes and florets were not covered with dried honeydew. Considering these differences from known sorghum parasites, this specimen remained unidentified.
The record of famine in Cameroon (1903–06), attributed to sorghum ergot disease, confirms that epiphytotics of C. africana occurred periodically even before the introduction of male-sterile A-lines of sorghum in the 1960s. Similarly, disease references and herbarium specimens document that C. africana was endemic in the sorghum-growing regions of Southern Africa well before the pathogen was recognized as a potential threat to production there.
Moreover, the history of specimen US0455138 from Liberia (Table 9) shows that it is possible to overlook sorghum ergot sclerotia, or sphacelia, due to their inconspicuous nature; the size, shape and reddish coloration mimics sorghum seed. The US phytosanitary control spotted Cerebella (considered a plant disease), but ergot was not recorded and no ergot specimen was deposited simultaneously. This historical example shows how inconspicuous C. africana infections can be. At low disease incidence and without copious honeydew exudation, the immature sphacelia and reddish sclerotia, which mimic the host seed in shape, could have been mistaken for sorghum grain, especially before the worldwide distribution of C. africana became apparent. Transfer with traded seed is a more probable explanation for the movement of the pathogen from Southern Africa to Brazil (confirmed by the presence of the same, but rare, haplotype in Hazyview, South Africa) than long-distance conidial transfer in air currents or even with contaminated clothing or footwear.
The existence of a population of C. africana with specialization towards a different andropogonoid host genus (Hyparrhenia), distant from Sorghum, suggests the limited potential for colonizing a wider spectrum of host grasses. Although the recent experiments with artificial inoculations of conidia of C. africana from sorghum to other grass genera were generally not successful (Reed et al., 2002; Montes-Belmont et al., 2003), pearl millet (P. glaucum), of tribe Paniceae, was artificially infected under high inoculum pressure (Futrell & Webster, 1966; Frederickson & Mantle, 1996; Muthusubramanian et al., 2005). Otherwise, only other Sorghum spp. have been implicated as important in the epidemiology (Ryley et al., 1996; Reed et al., 2002; Muthusubramanian et al., 2005).
The most widespread C. africana populations found in Southern Africa in 2000–02 are closely related. Representatives of the most widely distributed haplotype A, in South Africa, were first sampled in Matopos (Zimbabwe) in 1992, and were found in 1999 at the locations of Harare (Zimbabwe), Potchefstroom and Hazyview (both SA). However, this population was probably being replaced in Matopos during 2001–02 by a derived haplotype, L, differing by a single additional RAPD band.
Included in this study were C. africana isolates common to earlier variability studies, or made from the same collection trips (Table 1). Despite this, some disparity was noted between their placement in the trees obtained by the different methods. Komolong et al. (2002) wedged a clade of Ameroafrican isolates among the Australasian isolates: Australian isolate T 10765 appeared at the ‘ancestral’ position on the Ameroafrican clade, whereas isolate T 10763 was placed on the other side of the tree. As Komolong et al. (2002) stated that the observed differences were minor, it is surprising that the whole clade containing African and USA isolates did not appear outside the Australasian isolates from cluster 1, as a cluster of its own. In agreement with Komolong et al. (2002), the recent Indian population was very homogenous in both AFLP and RAPD analysis.
Tooley et al. (2002), using 130 AFLP markers, observed more dissimilarity between the isolates than in this study. Their values for diversity (HE) and Shannon index (I) were, in general, one order higher than those presented here. In their dendrogram, isolates from the same collections as isolates D1, E3 and F1 (in this study markedly differing in RAPD as well as AFLP pattern) were placed among the other African isolates, suggesting that their strains were not identical with the ones used here. Using primers MseI + C and EcoRI + AG, Tooley et al. (2000, 2002) observed an AFLP marker sized 111 bp in the American isolates. In the present analysis, the same marker occurred in all African haplotypes including the Hyparrhenia ones, but it was missing in the Eastern lineage. Therefore the loss of the marker in the Eastern lineage is a key discriminating event. AFLP (with 35 markers) was not able to discern as sharply as RAPD between African haplotype A and Ameroafrican haplotype W. With the three primer pairs there was no AFLP band specific for the latter group, whereas RAPD detected three.
The predominant clonality of C. africana populations in Africa, as confirmed by population genetics methods, is in agreement with the preferred theory of pathogen spread, through asexual primary conidia that are transferred with honeydew (over small distances in nature) and through secondary conidia which are airborne over moderate distances (Frederickson et al., 1991, 1993). The population structure in Southern Africa is probably the result of a widespread invasion of haplotype A that occurred in a single event or over several years, infection resembling the sweep of C. africana epiphytotics through the Americas in 1995–97. From this lineage, the origin of some haplotypes may be explained by a mutation (A to L), whereas in the case of B and C haplotypes a recombination event is possible (alleles shared with A, but private haplotypes).
The picture of C. africana population structure contrasts profoundly with that observed for Claviceps purpurea (Pažoutováet al., 2000b, 2002b). In this species, three isolated, habitat-specialized populations were found, one of them described as a subspecies (Duncan et al., 2002). However, within the most common population, G1 (fields and meadows), no two isolates had an identical RAPD (unpublished) or AFLP pattern (Pažoutováet al., 2002b). Moreover, there were no statistically supported clades or lineages. These are all typical features of a sexual panmictic population (for review see Taylor et al., 1999).
Very low Nem values between populations of C. africana and geographical areas indicate that, except for the original big invasion events, there has been almost no genetic exchange over 200–300 km. It may be speculated that a combination of favourable weather factors with an emerging vigorous clone is necessary to trigger epiphytotics like those encountered in the Americas and Australia in 1995–97 and 1996, respectively.
The unrelated haplotypes D–F might well be the remnants of the mosaic of local populations on which the clonal population of haplotype A was superimposed, according to the model of ‘epidemic population structure’ (Maynard Smith et al., 2000).
In the study of Komolong et al. (2002), most of the Australian isolates were closely related or identical, and belonged to the same lineage (denoted ‘Eastern’ in this study). This lineage was more closely related to recent African and American isolates than to the Australian ‘Cluster 2’ isolates. The authors suggest that the Cluster 2 isolates may have originated from previous invasion events that were unnoticed because they did not cause an epiphytotic of the same scale as the ‘Eastern’ lineage in 1996.
All American isolates, from Bolivia (1996), through Mexico and Puerto Rico (1998), to Brazil (2002), and including the isolates from the three locations at College Station, Texas, 1997, shared the same RAPD haplotype (W) with three bands unique to this lineage. This distinctive haplotype was found in two isolates from Hazyview (SA) in 1999 (Pažoutováet al., 2000a), and was not encountered in any other African region studied. Apparently the Hazyview isolates were the last remnant of the American invasion clone from 1995.
Tooley et al. (2002) hypothesized about two independent introductions of C. africana to the Americas based on two isolates, Cla42 and Cla38, from Nebraska, 1997 appearing far apart on his dendrogram. There are several arguments against two independent invasion events. First, with results similar to those presented in this study, and in contrast to Tooley et al. (2002), the data of Komolong et al. (2002) show Cla5 (College Station, TX, 1997) and Cla14 (Isabela, Puerto Rico, 1998) as sisters on the same clade. Furthermore, official documentation and publication records began in 1995 and were especially strong during 1996–97, demonstrating that scientists and, indeed, the entire seed industry were actively looking for ergot across the western hemisphere (for review see Bandyopadhyay et al., 1998). Therefore the timing of the initial disease outbreaks, and the subsequent pattern and incidence of disease spread, are precisely known. With the molecular data from samples collected sequentially with spread and concurrently with disease distribution, the facts indicate the redistribution of only one clone, haplotype W, which first arrived in Brazil in 1995 and subsequently gave rise to infection in all other areas.
S. Pažoutová is grateful to Dr John Leslie (Department of Plant Pathology, Kansas State University) for permitting her to conduct AFLP analyses in his laboratory in 2000. Thanks are due to Dr N. McLaren for assisting D.E. Frederickson with sampling of sorghum ergot in the Republic of South Africa. INTSORMIL and the Czech Institutional Research Concept No. AV0Z5020903 provided the funding for this study.
- 2000. Multilocus 1·2. London, UK: Imperial College, Department of Biology. , ,
- 1998. Ergot: a new disease threat to sorghum in the Americas and Australia. Plant Disease 82, 356–67. , , , , ,
- 1986. A quasi-equilibrium theory of the distribution of rare alleles in a subdivided population. Heredity 56, 409–15. , ,
- 1989. Nordiska-Cameroun, crise et peuplement. Rouen, France: Université de Rouen, PhD thesis. ,
- 1980. Multilocus structure of natural populations of Hordeum spontaneum. Genetics 96, 523–36. , , ,
- 1950. The South African fungi and lichens. Bothalia 5, 1–1094. ,
- 2002. Claviceps purpurea var. spartinae var. nov. an ergot adapted to the aquatic environment. Mycotaxon 81, 11–25. , , , , ,
- 1975. A simple test for the possible simultaneous evolutionary divergence of two amino acid positions. Taxon 24, 609–13. , ,
- 1996. Pearl millet as an alternate host of the sorghum ergot pathogen, Claviceps africana. International Sorghum and Millet Newsletter 37, 83–5. , ,
- 1991. Claviceps africana sp. nov., the distinctive ergot pathogen of sorghum in Africa. Mycological Research 95, 1101–7. , , ,
- 1993. Windborne spread of ergot disease (Claviceps africana) in sorghum A-lines in Zimbabwe. Plant Pathology 42, 368–77. , , ,
- 1966. Host range and epidemiology of the sorghum ergot organism. Plant Disease Reporter 50, 828–31. , ,
- 2000. lian 3·0: detecting linkage disequilibrium in multilocus data. Bioinformatics 16, 847–8. , ,
- 2002. Identity and genetic diversity of the sorghum ergot pathogen in Australia. Australian Journal of Agricultural Research 53, 621–8. , , , ,
- 1952. Studies on ergot. Brisbane, Australia: University of Queensland, PhD thesis. ,
- 1954. The origin and differentiation of Claviceps species. University of Queensland Papers (Botany) 3, 61–8. ,
- 1955. The genus Cerebella. Mycological Papers 61, 1–18. ,
- 1972. The apportionment of human diversity. Evolutionary Biology 6, 381–98. ,
- 1964. Use of the honeydew state in the identification of ergot species. Transactions of the British Mycological Society 47, 205–13. ,
- 1985. The occurrence of ergot fungi (Clavicipitales) on the genus Hyparrhenia in southern Africa. Botanical Journal of the Linnean Society 91, 491–501. ,
- 1999. The detection of recombination from sequence data. Genetics 153, 1021–7. ,
- 2000. Population structure and evolutionary dynamics of pathogenic bacteria. Bioessays 22, 1115–22. , , ,
- 1923. Report of the Mycologist. Annual Report of Kenya. Nairobi, Kenya: Department of Agriculture, 81–5. ,
- 1925. Report of the Mycologist. Annual Report of Kenya. Nairobi, Kenya: Department of Agriculture, 145. ,
- 1997. Changes in pollen viability and concomitant increase in the incidence of sorghum ergot with flowering date and implications in selection for escape resistance. Journal of Phytopathology 145, 261–5. ,
- 1998. Use of weather variables to quantify sorghum ergot potential in South Africa. Plant Disease 82, 26–9. , ,
- 1990. Relationship between climatic variables during flowering of sorghum and the incidence of sugary disease caused by Sphacelia sorghi. Journal of Phytopathology 130, 82–8. , ,
- 1992. Pre-flowering low temperature predisposition of sorghum to sugary disease (Claviceps africana). Journal of Phytopathology 135, 328–30. , ,
- 1997. Tools for population genetic analysis (tfpga) 1·3: A Windows program for the analysis of allozyme and molecular population genetic data. http://www.marksgeneticsoftware.net ,
- 1991. New records of ergot of sorghum caused by Sphacelia sorghi in southern Africa. Plant Disease 75, 215. , , , , , , , , ,
- 2003. Alternate hosts of Claviceps africana Frederickson, Mantle and de Millano, causal agent of sorghum ergot in the State of Morelos, Mexico. Revista Mexicana de Fitopatología 21, 63–6. , , ,
- 2005. Inoculated host range and effect of host on morphology and size of macroconidia produced by Claviceps africana and Claviceps sorghi. Journal of Phytopathology 153, 1–4. , , , ,
- 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89, 583–90. ,
- 1987. Molecular Evolutionary Genetics. New York, USA: Columbia University Press. ,
- 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proceedings of the National Academy of Sciences, USA 76, 5269–73. , ,
- 1999. Experiences with ergot of sorghum in South Africa. In: CaselaCR, DahlbergJA, eds. Proceedings of the Global Conference on Ergot of Sorghum, 1997. Publication no. 99–1. Lincoln, NB, USA: INTSORMIL, 36–40. , ,
- 2000a. Relations among sorghum ergot isolates from the Americas, Africa, India, and Australia. Plant Disease 84, 437–42. , , , , ,
- 2000b. Claviceps purpurea populations, chemoraces and habitat specialization. Applied and Environmental Microbiology 66, 5419–25. , , , , ,
- 2002a. Heteropogon triticeus– a new host for Claviceps sorghi in India. Journal of Phytopathology 150, 1–4. , , , ,
- 2002b. Specialised population of Claviceps purpurea from salt marsh Spartina species. Mycological Research 106, 210–4. , , , ,
- 1997. Construction of evolutionary distance trees with treecon for Windows: accounting for variation in nucleotide substitution rate among sites. Computers and Applied Biosciences 13, 227–30. , ,
- 1995a. An exact test for population differentiation. Evolution 49, 1280–3. , ,
- 1995b. genepop, version 1·2: a population genetics software for exact tests and ecumenicism. Journal of Heredity 86, 248–9. , ,
- 2002. Analysis of resistance to ergot in sorghum and potential alternate hosts. Crop Science 42, 1135–8. , , , ,
- 1996. First report in the Americas of sorghum ergot disease, caused by a pathogen diagnosed as Claviceps africana. Plant Disease 80, 463. , , ,
- 1996. Ergot on Sorghum spp. in Australia. Australian Plant Pathology 25, 214. , , , , ,
- 2000. Arlequin: A Software for Population Genetic Data. Geneva, Switzerland: University of Geneva, Genetics and Biometry Laboratory. , , ,
- 1995. Biometry, 3rd edn. New York, USA: W.H. Freeman. , ,
- 1999. The evolutionary biology and population genetics underlying fungal strain typing. Clinical Microbiology Review 12, 126–46. , , , ,
- 1980. Pollination effects on pearl millet ergot. Phytopathology 70, 80–8. , ,
- 1991. Are eukaryotic microorganisms clonal or sexual? A population genetics vantage. Proceedings of the National Academy of Sciences, USA 88, 5129–33. , , , , , , ,
- 2000. Use of random amplified microsatellites and AFLPs in assessing phenotypic diversity among isolates of Claviceps africana and other Claviceps species. Phytopathology 90, 1126–30. , , , ,
- 2002. AFLP comparisons among Claviceps africana isolates from the United States, Mexico, Africa, Australia, India, and Japan. Plant Disease 86, 1247–52. , , , ,
- 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research 23, 4407–14. , , , , , , , , , ,
- 1949. A List of Plant Diseases in Tanganyika Territory. Mycological Paper no. 26. Kew, UK: Commonwealth Mycological Institute. , ,
- 2003. Effect of climate variability on event frequency of sorghum ergot in Australia. Australian Journal of Agricultural Research 54, 599–611. , , ,
- 1984. Estimating F-statistics for the analysis of population structure. Evolution 38, 1358–70. , ,
- 1926. Report of the Mycologist. Annual Report of Kenya. Nairobi, Kenya: Department of Agriculture. ,
- 1978. Evolution and the Genetics of Populations, Vol. 4: Variability Within and Among Natural Populations. Chicago, IL, USA: University of Chicago Press. ,
- 2000. Popgene32, Microsoft Windows-based freeware for population genetic analysis, version 1·32. Edmonton, Canada: University of Alberta, Molecular Biology and. Biotechnology Centre. , , , , ,
- 2000. Isozyme and AFLP from Cephalosporium maydis in Egypt. Phytoparasitica 28, 121–30. , , , ,