Fig rust caused by Phakopsora nishidana in South Africa

Summary. Fig rust, caused by Cerotelium fici , was first recorded in South Africa in 1927. Recent observations have revealed high incidence of rust and untimely defoliation of fig trees ( Ficus carica ) in residential gardens and commercial orchards. Using phylogenetic analysis, the causal organism of a fig rust isolate (PREM63073) collected in 2020 was confirmed as Phakopsora nishidana . Inoculation and microscope studies showed that mulberry plants were immune to P. nishidana isolate PREM63073. Infection of fig leaves occurred through stomata on the abaxial leaf surfaces. Very long germ tubes were observed for P. nishidana , often with no clear contact with the leaf surfaces and an apparent lack of directional growth towards stomata. Inoculated plants from 15 fig cultivars varied in their severity of leaf infection, whereas fruit of the cultivar Kadota developed reddish-brown blemishes without sporulation. Currently, C. fici and P. nishidana are recognised as occurring on F. carica in South Africa. This suggests a need to resolve the worldwide distribution and identity of the rust species involved.


INTRODUCTION
Ficus belongs to the Moraceae, which contains 60 genera and at least 1400 species (Gerber, 2010). The edible fig, Ficus carica L., one of approximately 750 Ficus species (Lötter, 2014), is native to the Middle East and Mediterranean regions and thrives where winters are mild and summers are hot and dry (Verga and Nelson, 2014) According to Karsten (1951), Godée Molsbergen mentioned figs as one of the many cultivated plants introduced to the South African Cape region between 1652 and 1662. Initially, fig production was restricted to individual trees or single rows or avenues on farms and residential gardens. In the absence of commercial production during the 18th and 19th centuries, fresh figs were used mostly for local consumption, but also as dried products and different types of preserve (Lötter, 2014). In 1902/03, the Cape Department of Agriculture imported 140 'Smyrna' and 'Capri' trees of different cultivars (e.g., Calimyrna, Kassaba, Bardajie,and Capri No. 1,No. 2 and No. 3), providing momentum to the local industry. Once the trees were established, 'Capri' cuttings with fruit containing Blastophaga wasps were imported under carefully controlled conditions for the insect to serve as a pollinating vector. South African fig production, located mainly in the Western Cape Province, varied considerably over the years (Lötter, 2014). Over the past 50 years attempts to revitalize the fig industry were launched by establishing experimental blocks and importing additional cultivars from California and France. Doidge (1927;1950) listed fig rust specimens collected from Mozambique, Zimbabwe, and the Eastern Cape, Free State, KwaZulu-Natal, North West, and Western Cape provinces in SA as belonging to Cerotelium fici (Castagne) Arthur. In addition, Verwoerd (1929) reported the occurrence of C. fici in the Western Cape, specifically in Wynberg, Paarl and Stellenbosch, and also from drier localities such as Barrydale, Ladismith and Oudtshoorn.
Fig rust symptoms are first visible as flecks on upper leaf surfaces, followed by the appearance of sporulating pustules on the lower surfaces. Pustules can occur over entire leaf blades, but are usually clustered in areas where moisture accumulates (McKenzie, 2013). Eventually, the spots on the upper surface of each affected leaf become brown and angular and may coalesce to form larger necrotic areas, leading to defoliation under conditions of severe infection. Fruit infections are visible as blister-like rust pustules (Verga and Nelson, 2014). Mulberry (Morus spp.) has also been reported as a host for C. fici, although Laundon and Rainbow (1971) doubted the conspecificity of the rust fungi on fig and mulberry.
In 2020, a fig tree heavily infected with rust was observed in a residential garden (Figure 1). The vis-ible impact of the disease on tree health prompted an investigation to: (i) confirm the identity of the causal organism; (ii) understand the infection process; and (iii) evaluate methods to determine the host range and response of fig cultivars to the rust isolate under controlled conditions.
The amplicon was cloned into the pGem-T® Easy plasmid vector (Promega Corporation) and transferred into Escherichia coli JM109 competent cells. Cloned inserts of four recombinant plasmids were sequenced with the Rust2Inv, primer LR6, and internal primers LR0R (Moncalvo et al., 1995) and LR3 (Vilgalys and Hester, 1990), using the BigDye TM Terminator v 3.01 Sequencing kit (Thermo Fisher Scientific). Sequenced products were separated on a 3130x1 Genetic Analyzer (Applied Biosystems), using the StdSeq50_POP7 sequencing run module. Once all ambiguous nucleotides were corrected, the four overlapping sequences of each recombinant clone were assembled into a single contiguous sequence, using the online CAP3 Sequence Assembly Program (Huang and Madan, 1999). The final consensus sequence was deposited on GenBank (accession No. MZ047090).
The two allelic sequences were aligned with selected reference sequences (Table 1), using the online MAFFT web interface (Katoh et al., 2019). Helicobasidium longisporum was used as the outgroup (Maier et al., 2016). After trimming the 5' and 3' extending sequences, the resulting 1036 bp ITS2-28S rRNA sequences were used to determine the appropriate nucleotide substitution model with the Akaike Information Criterion (AIC) within jModelTest v 2.1.1 (Darriba et al., 2012;Guindon and Gascuel, 2003). The suggested TIM2 + G + I (n = 5) nucleotide substitution model was replaced with the GTR + G + I (n = 5) model (Lecocq et al., 2013) for Bayesian inference (BI) analysis in MrBayes v 3.1.2 (Huelsenbeck and Ronquist, 2001;Ronquist and Huelsenbeck, 2003). BI analysis was started from a random tree using four Markov Chain Monte Carlo (MCMC) chains. The search was limited to 5,000,000 searches, where every 500 th generation was sampled. The average standard deviation of split frequencies was examined and the analysis stopped at a value below 0.01. The first 1000 trees were discarded as burnin before analysis.
Maximum parsimony (MP) analysis was carried out in PAUP* v. 4.0b10 (Swofford, 2003), where all characters were weighted equally. The heuristic search was carried out with 1000 addition-sequence replicates, with tree bisection and reconnection (TBR) branch swopping. Ten trees per replicate were saved. Bootstrap support for proposed branches was evaluated with 1000 replicates with 100 random addition-sequence replicates and TBR branch swopping.

Spore and pustule morphology
The widths and lengths of 60 urediniospores of isolate PREM63073 were measured with an Olympus BX53 light microscope, fitted with a DP72 digital camera for image capturing with Analysis LS Research version 2.2 software (Olympus Soft Imaging System). Free-hand cross-sections of pustules (uredinia and telia) were made after soaking small pieces of dried leaves in hot water for a minimum of 30 min. Spores and pustules were examined with either a Zeiss Axioscope or a Nikon E600 microscope.  bags filled with a mulch mixture, and acclimatized for 3 weeks pre-inoculation in a greenhouse cubicle with a temperature range of 18 to 25°C. The plants were watered every second day with reverse osmosis water, and were each fertilised once a week with 100 mL of a 0.2% (w/v) Multifeed-Classic water-soluble fertilizer [Effekto®, NPK analysis 19:8:16 (43)]. Cuttings representative of Morus alba L., including common or white mulberry, as well as weeping mulberry, were sampled a few hours before inoculation from confirmed sources. These cuttings were placed in 10 L containers filled with reverse osmosis water before and after inoculation. Urediniospores of isolate PREM63073, kept at -80°C, were used in all experiments. Three weeks before inoculation, the urediniospores were increased through the inoculation of leaves of the fig cultivar Kadota. Procedures described by Boshoff et al. (2020) were used to heat-shock the spores and for inoculation of plants and subsequent incubation. After 3 weeks, urediniospores were collected from the Kadota leaves into size 00 gelatine capsules, by connecting an air vacuum to a cyclone spore collector (Pretorius et al., 2019).

Host infection studies
The response of fig and mulberry plants, including the mulberry cuttings, to the local rust isolate, was studied in two independent trials. During the first trial, a urediniospore concentration of 98 × 10 4 spores mL -1 in 0.3 mL Soltrol® 130 isoparaffinic oil, was applied per leaf. Two leaves were inoculated per cultivar/cutting, one inoculated on the abaxial surface and the other on the adaxial surface. In the second trial, one leaf per cultivar/cutting was inoculated on the abaxial surface, using a urediniospore concentration of 87.7 × 10 4 spores mL -1 in 0.3 mL Soltrol® 130 oil, as described above. During the second trial, replicate young developing figs from the cultivar Kadota were also inoculated using the above urediniospore concentration. Inoculated leaves on plants, young figs and cuttings were regularly observed for development of first symptoms of infection. Three weeks after inoculation, the number of rust pustules per cm 2 was counted on the abaxial leaf surfaces. This was achieved by placing a rectangular template with a 1 cm 2 opening at five random positions on each assessed leaf.
Analysis of variance (ANOVA; α = 0.05) of host response data was performed for the number of rust pustules counted per cm 2 , using 'base' R functions, where the ANOVA model 'pustules ~ cultivar + trial' was applied (N = 150). Means were separated on the Minimum Significant Difference (MSD) test, using the LSD.test function from the 'agricolae' package, with an adjusted Bonferroni P-value to account for potential family-wise error rates (De Mendiburu, 2020). Data processing and analyses were performed with R version 4.0.2 (R Core Team, 2020) within R Studio version 1.2.5042 (R Studio Team, 2020). Data exploration, wrangling and visualisation were conducted using the 'Tidyverse' package (Wickham et al., 2019). Data scripting was conducted in rmarkdown (Allaire et al., 2020).
Fluorescence and scanning electron microscopy (SEM) were used to describe the host penetration process by the pathogen. Material for microscopy included newly sprouted leaves from the susceptible fig cultivars Cape White and Kadota, as well as young developing Kadota figs. Young leaves from the mulberry cultivar, Queensland Red, and freshly collected cuttings from a weeping mulberry plant, were included for fluorescence microscopy. The abaxial surfaces of three leaves per plant/cutting were each sprayed with 0.8 mL of urediniospore suspension (±1 mg mL -1 urediniospores in Soltrol® 130 oil), whereas dry spores were applied with a brush to young fruit. Post-inoculation treatment, dew chamber incubation and greenhouse conditions were as described above.

Fluorescence microscopy
Leaf and fig peel segments were sampled at 48 and 96 h (hpi), and 16 d post-inoculation (dpi), and were cut into 10 mm segments. Half of the samples were prepared using a modified method of Rohringer et al. (1977) as described by Moldenhauer et al. (2006), and stained leaf segments were stored in 50% (v/v) glycerol containing a trace of lactophenol to prevent deterioration and drying of fungal material.
For the segments stained with Uvitex 2B, the blue wavelength epifluorescence cube with an excitation filter of 330-385 nm and a barrier filter of 420 nm showed fluorescence of stained fungal tissue. Observations of WGA-FITC stained segments were made with the ultraviolet wavelength (WU) epifluorescence cube, with an excitation filter of 450-480 nm and a barrier filter of 515 nm. The microscope was fitted with a CC12 digital camera for image capturing with Analysis LS Research version 2.2 software (Olympus Soft Imaging System).

Scanning electron microscopy
Leaf and fig peel segments were sampled at 48 and 96 hpi, and 16 dpi, and were cut into 5 mm segments. Samples were fixed according to the protocol of Glauert (1974). The dried samples were directly mounted on 12.2 mm diam. metal stubs (Cambridge pin type) using double-sided carbon tape, for observations of fruit or abaxial leaf surfaces. The mounted segments were coated with gold (± 60 nm thickness) in a sputter coater (Bio-Rad), and the specimens examined with a JSM-7800F Extreme-resolution Analytical Field Emission SEM.

Molecular identification of the fig rust isolate
Sequencing of multiple recombinant clones identified two ITS2-28S rRNA allelic variants of isolate PREM63073, which differed with four nucleotides within the ITS2 region over the 1054 bp amplicon. A  Sister to this, was a clade that grouped both allelic variants of PREM63073 with one C. fici (likely incorrectly identified) and six other Phakopsora isolates. This clade was divided into two sub-clades (83% BS; 0.98 PP). The first sub-clade contained both PREM63073 variants, and three P. nishidana accessions that were all collected from F. carica trees from, respectively, SA, the USA and Mexico. The second sub-clade contained four P. myrtacearum isolates, all collected from Eucalyptus trees in eastern Africa.
The other five C. fici isolates grouped as a separate sub-clade (68% BS; 0.99 PP), within a larger clade (64% BS; 0.94 PP) containing P. jatrophicola (collected from Jatropha sp.), P. pachyrhizi (Glycine max and Desmodium sp.), P. meibomiae (Aeschynomene sp.) and P. tecta (Commelina sp.). The first of these five C. fici isolates (KP753385) formed part of a study on P. myrtacearum on eucalyptus trees in southern and eastern Africa, where the pathogen origin was indicated as a Ficus sp. in Australia (Maier et al., 2016). The other four were direct submissions to GenBank as unpublished studies. The first two of these accessions (MH047209 and MH047210) were part of a study of Uredo morifolia on mulberry in Australia, where these isolates were collected from F. coronulata. The other two C. fici accessions (MK135779 and MK135780) were collected from Ficus spp. in Pakistan.

Infection studies
No symptoms or signs developed on the inoculated leaves of mulberry cultivar Queensland Red or on the cuttings from the common or weeping mulberry plants.
This apparent immunity of the tested mulberry plants was confirmed by fluorescence microscopy. Observations at either 48 or 96 hpi showed random distribution of appressoria, occasionally over stomata, on the mulberry leaves. Here, abortion of substomatal vesicles without development of haustorium mother cells (HMCs) (Figure 4, A) or non-penetrating appressoria (Figure 4, B), were observed.
Fig leaves inoculated on the adaxial surfaces did not develop meaningful infections (zero to trace infection), and the leaves remained mostly symptomless. The first signs of infection became visible as minute reddish-brown flecks for leaves inoculated on the abaxial surfaces, 7 dpi. The latent period, indicated by the first signs of sporulation, varied among cultivars, and was 10 dpi for Brunswick and Noire de Caromb. This was followed by 11 dpi for Tiger, Black Mission, Dalmatie, Ronde de Bordeaux, Col de Dame Noir, Parisian, Deanne, Adamsvy and Cape White at 11 dpi; 12 dpi for Kadota and Cape Brown; 13 dpi for Black Genoa; and 14 dpi for White Genoa. The ANOVA indicated no statistically significant differences (F = 3.7, P = 0.06) between the two experimental trial replicates for mean numbers of pustules per cm 2 . However, significant differences were detected between the host cultivars (F = 156.3, P < 0.05). Cape White, Noire de Caromb (mean >20 pustules per cm 2 ) and Tiger (mean > 15 pustules per cm 2 ) were the most severely affected cultivars (Figure 5). While the Cape White and Noire de Caromb cultivars did not differ (P > 0.05) from one another, they were different from Tiger. The remaining cultivars, except White Genoa, produced mean numbers of pustules per cm 2 between 5.0 and 13.6, with three clusters of responses indicated by mean separation. White Genoa responded the least to inoculation, producing a mean of 3.2 pustules per cm 2 . Rust severity (27 dpi) of infected leaves for cultivars Noire de Caromb ( Figure  6, A), Brunswick (Figure 6, B), Adamsvy (Figure 6, C), Black Genoa (Figure 6, D), and 50 dpi for Brunswick on the adaxial (Figure 7, A) and abaxial (Figure 7, B) leaf surfaces, indicated severe disease on some cultivars. Inoculated figs did not show typical rust signs or symptoms. However, as the fruit matured, reddishbrown blemishes became visible on the peels (Figure 8, A). Although germinated urediniospores were visible with SEM at 48 hpi on the skin of young Kadota fruit (Figure 8, B), sometimes close to stomata, no appressoria were observed. At 96 hpi, the germ tubes were collapsed with no signs of infections. Instead, blister-like structures, usually associated with host stomata in the urediniospore treated areas, but not necessarily with presence of rust spores, were observed at 96 hpi (Figure Light microscopy showed that leaves of Cape White and Kadota all had infection structures, typical of those formed on leaf surfaces by rust fungi during the first phases of infection cycles. At 48 hpi, these structures included germinated urediniospores, germ tubes, and appressoria forming over stomata on the abaxial leaf surfaces (Figure 9, A and B). As each appressorium matured, it was delimited from the germ tube by a septum (Figure 9, B and C) followed by collapse of the germ tube. Most of the appressoria were collapsed on top of stomata at 96 hpi (Figure 9, C). The P. nishidana germ tubes were very long, often extending from one trichrome to another with no apparent contact with the leaf surface (Figure 9, D). Formation of appressoria was not necessarily on the first encountered stomata (Figure 9, E), and appeared to be random. With fluorescence microscopy (not shown), only a few HMCs (mostly one or two) were visible at 48 hpi, whereas at 96 hpi, small colonies (approx. 30 HMC's) were observed. At 14 dpi, established colonies filled with urediniospores were common (Figure 9, F).  (Doidge, 1927(Doidge, , 1950. The phylogenetic analysis of the present study showed that both these species belong to the Phakopsoraceae, as currently delimited (Aime and McTaggart, 2020). Phakopsora nishidana is in a distinct clade to that containing the type species, P. pachyrhizi Syd. & P. Syd. (Aime and McTaggart, 2020), and therefore should be assigned to a new genus. This has not been conducted in the present study, as more species need to be included in analyses before generic limits within the Phakopsoraceae can be accurately assessed. Cerotelium fici is, in the present analysis, a sister clade to Phakopsora sensu stricto, and  may therefore be transferred to this genus depending on results of additional analyses.
The application of the names Cerotelium fici and Phakospora nishidana has been confused, with some workers applying the C. fici for the widely distributed fig rust pathogen (e.g. Laundon and Rainbow, 1971;McKenzie, 1986), whereas others have applied P. nishidana (e.g. Buritica, 1999;Hennen et al., 2005;De Carvalho et al., 2006). This resulted from differing interpretations of the species to which the first described spore stage, Uredo fici Castagne, belongs. Here, this is assigned to C. fici. The paraphyses of P. nishidana were described as both peripheral and intermixed with the urediniospores (Ito and Homma, 1938). In contrast, the uredinia of U. fici are surrounded by paraphyses (malupa-like) (Arthur, 1906), as described for C. fici by Butler (1914). In addi-   Otani ex S. Ito & Muray., not P. nishidana. However, P. nishidana was not recorded from F. carica (Ito and Murayama, 1949), the type host of U. fici. The uredinial morphology of P. fici-erectae should be re-examined and contrasted to that of C. fici, as this was beyond the scope of the present study.
The correct synonomy for the two species considered, is as follows: Note: under the previous rules of Nomenclature, a teleomorph name could not be based on an anamorph type specimen, and therefore this species was correctly cited as Cerotelium fici (E.J. Butler) Arthur, as Butler's K. fici was then the correct basionym. However, since the principle of 'one fungus one name' has been applied, the first correctly published name becomes the basionym, even if for an anamorph. Besides these species, another, Uredo sawadae S. Ito (Farr and Rossman, 2021) has been described from F. carica. Other rust species described from various Ficus species, other than F. carica, include Phakopsora ficielasticae T.S. Ramakr., Cerotelium ficicola Buriticá & J.F. Hennen, Crossopsora fici Arthur & Cummins (Farr and Rossman, 2021) and various Uredo species that must be assigned to one of these listed telial species or as yet undescribed species.
Cerotelium fici often occurs wherever cultivated figs are produced (Laundon and Rainbow, 1971;McKenzie, 1986;Latinović et al., 2015). Except for the reports of Doidge (1927;1950), Verwoerd (1929) and Lötter (2014), no published information on fig rust in SA could be found. As the description provided by Doidge (1927) corresponds to other accounts of C. fici (Laundon and Rainbow, 1971;McKenzie, 1986McKenzie, , 2013, the historical context and urediniospore morphology suggested that the Vermont rust isolate could be C. fici. Phylogenetic analysis, however, clearly differentiated between C. fici and P. nishidana. The close association of molecular data of the local fig rust examined with two P. nishidana accessions confirmed that isolate PREM63073 was P. nishidana and not C. fici. The placement of the Mexican C. fici isolate UACH-107 (GenBank accession MF580676; Solano-Báez et al., 2017) within this subclade indicated that it was incorrectly identified, due to a lack of adequate reference sequences. The morphology described for this isolate (Solano-Báez et al., 2017) is consistent with that of the uredinial state of P. nishidana, though the urediniospores were slightly larger (23-27 × 16-19 µm) than originally described for this species (18-24 × 16-18 µm;Ito and Homma, 1938). One of the reference sequences for P. nishidana (KY764080) was of a specimen that originate from Mexico (BPI 910197), confirming the presence of this species in that country. Furthermore, a general assumption that fig rust is caused by C. fici, as well as morphological similarities between the urediniospores of C. fici and P. nishidana, may prevent clear visual distinction between the two species.
Teliospore morphology is distinct between C. fici and P. nishidana (Ito and Homma, 1938), with the telia of C. fici being typical of Cerotelium, as erumpent and producing teliospores in short chains which are not laterally adherent. The telia of P. nishidana are not erumpent, with the teliospores forming crusts of laterally adherent teliospores. Telia are seldom produced by C. fici, having only been recorded three times (Butler, 1914;Patil and Thirumalachar, 1971;Huseyin and Selcuk, 2004), whereas telia are readily produced by P. nishidana (Ito and Homma, 1938). The urediniospores of these two fungi are morphologically almost identical, being hyaline, thin-walled and with inconspicuous germ pores in both species. They differ in that they are slightly smaller in P. nishidana (18-24 × 16-18 µm;Ito and Homma 1938: typically 19-26 × 16-19 µm, though occasionally larger or smaller in the present study), compared to C. fici (20-35 × 14-25 µm: Butler, 1914;McKenzie, 1986;2013;Huseyin and Selcuk, 2004;Latinović et al., 2015). However, as stated above, the uredinia in P. nishidana are uredo-like, with paraphyses intermixed with the urediniospores, whereas those of C. fici are malupalike with only peripheral paraphyses (Butler, 1914;Doidge, 1927;Ellis, 2020). As the present study collection had Phakopsora-like telia present, and the uredinia had paraphyses intermixed through these pustules, this confirmed the rust's identity as P. nishidana, despite some of the urediniospores being larger than previously described for this species.
The taxonomy of rust fungi on figs remains confused, and there is need to fully resolve the species involved and their distribution, both within and beyond fig production regions. Phakopsora nishidana has been recorded from eastern Asia and the Americas (Hennen et al., 2005;Solano-Báez et al., 2017;Farr and Rossman, 2021), and there is a possibility that some of the records of C. fici from southern Africa, as well as elsewhere in Africa, may also be P. nishidana. The species with Phakopsora-like telia described from Ficus hosts, including P. nishidana, P. fici-erectae and P. fici-elasticae, as well as U. sawadae, need to be compared by phylogenetic analyses. The description of the fig rust identified as P. nishidana from South America (Buritica, 1999;Hennen et al., 2005) fits the description of P. fici-erectae better than that for P. nishidana. It is possible that several species are involved, but are not recognised, as the urediniospore morphology for all is almost identical.
In SA, the Western Cape Province, specifically around the towns of Porterville, Malmesbury, Paarl and Wellington, and Napier in the Overberg, is considered as the most important fig producing region. The production conditions in this province are diverse and vary from more rust-conducive areas, with regular conditions of high humidity along the coastal regions, to semi-arid and less disease-prone areas such as Worcester, Ladysmith, Oudtshoorn and Prince Albert. Although rust regularly occurs towards the end of the production season in fig orchards in the Overberg, infection levels are generally low and without apparent economic impacts. In April 2021, however, rust incidence in orchards near Napier was high, with several trees defoliating due to rust infections. Similar to the observations of McKenzie (2013), rust pustules were typically clustered alongside leaf veins or on leaf sections exposed to moisture accumulation and retention. Seasonal variation in severity and distribution of rusts is influenced by inoculum levels, environment, and cultivar susceptibility. The mulberry plants tested in the present study were all highly resistant to rust infections, indicating that the rust fungus found elsewhere on this host may be different from P. nishidana which infects edible fig plants.
Phakospora nishidana is known as a pathogen only of several species of Ficus (Farr and Rossman, 2021), whereas C. fici is recorded from species of Broussonetia, Ficus, Maclura and Morus (Moraceae) (McKenzie, 1986;Farr and Rossman, 2021). Inhibition of colony establishment at an early stage, as seen in both mulberries, was classified as abortive penetration according to Kochman and Brown (1975). In F. carica, infection occurs after formation of appressoria over stomata on abaxial leaf surfaces of the hypostomatic leaves. Directional germ tube growth was not observed for P. nishidana, but it was evident that trichomes on lower leaf surfaces interfered with germ tube extension. Infection of lower leaf surfaces, and thick upper epidermis with underlying leaf anatomical structures, may explain the predominant colonization and sporulation on the abaxial surfaces.
Selection of cultivars tested in the present study was influenced by the availability of trees at local nurseries. Although none of the cultivars was immune to rust infections, differences were observed in the severity of infections, which may impact disease development and economic losses. Parisian, the dominant purple fig cultivar grown in SA, which accounts for about 75% of fresh fig production, and Ronde de Bordeaux, were intermediate in their rust responses. Deanna is a yellow peelcoloured cultivar grown especially in the North West Province. It is early and productive, with large fruit for the local market, and was less susceptible to rust in the present study. Mean rust severity on Black Genoa, which is grown in many countries under names such as San Piero, California Brown Turkey, Roxo de Valinhos, Black Jack, and Negro Largo, was significantly less than on most other cultivars. In the present greenhouse study, cultivar responses were assessed as the number of pustules per unit leaf area. However, the severity scale developed by Da Silva et al. (2019) provided a useful resource should rust assessments be expanded to field trials. The present results have shown that routine methodology to study rust pathosystems of field crops is equally applicable for fig rust research, and has potential for further application in horticultural systems. Rogovski Czaja et al. (2021)  Control of diseases on figs is hampered by the lack of coordinated research including data on the occurrence and economic importance of a disease such as rust. Furthermore, fresh figs produced for export must adhere to pesticide residue limits which are strictly imposed by importing countries. At present, SA is excluded from exporting to the biggest markets of the USA and China until regulatory protocols have been established, due to the phytosanitary requirements for these countries. These regulations and the relatively small area planted with commercial figs in SA have contributed to the scarcity of relevant published informa-tion on pests and pathogens on fig, as well as almost non-existence of chemicals registered to for their control. Advice on chemical control of fig rust to commercial producers and residential fig growers is currently limited due to a lack of registered products in SA. Kanguard 940™, labelled as a contact fungicide and containing plant organic acids as active ingredients, is the only disease control product registered on figs in SA (https:// www.agri-intel.com; http://www.kannar.co.za; accessed 26 October 2020). Spray programmes involving this product require preventative application at 5 d intervals and good coverage of plants. However, the efficacy of this product has not been confirmed against P. nishidana and requires further evaluation along with other fig rust management options.