Biology and Epidemiology of Venturia Species Affecting Fruit Crops: A Review

The fungal genus Venturia Sacc. (anamorph Fusicladium Bonord.) includes plant pathogens that cause substantial economic damage to fruit crops worldwide. Although Venturia inaequalis is considered a model species in plant pathology, other Venturia spp. also cause scab on other fruit trees. Relative to the substantial research that has been conducted on V. inaequalis and apple scab, little research has been conducted on Venturia spp. affecting other fruit trees. In this review, the main characteristics of plant-pathogenic species of Venturia are discussed with special attention to V. inaequalis affecting apple, V. pyrina affecting European pear, V. nashicola affecting Asian pear, V. carpophila affecting peach and almond, Fusicladium oleagineum affecting olive, F. effusum affecting pecan, and F. eriobotryae affecting loquat. This review has two main objectives: (i) to identify the main gaps in our knowledge regarding the biology and epidemiology of Venturia spp. affecting fruit trees; and (ii) to identify similarities and differences among these Venturia spp. in order to improve disease management. A thorough review has been conducted of studies regarding the phylogenetic relationships, host ranges, biologies, and epidemiologies of Venturia spp. A multiple correspondence analysis (CA) has also been performed on the main epidemiological components of these Venturia spp. CA separated the Venturia spp. into two main groups, according to their epidemiological behavior: the first group included V. inaequalis, V. pyrina, V. nashicola, and V. carpophila, the second F. oleagineum and F. eriobotryae, with F. effusum having an intermediate position. This review shows that Venturia spp. affecting fruit trees are highly host-specific, and that important gaps in understanding the life cycle exist for some species, including V. pyrina; gaps include pseudothecia formation, ascospore and conidia germination, and mycelial growth. Considering the epidemiological information reviewed, this paper shows that the use of Mills tables to predict infection periods should be avoided for Venturia spp. other than V. inaequalis.

The fungal genus Venturia Sacc. (anamorph Fusicladium Bonord.) includes plant pathogens that cause substantial economic damage to fruit crops worldwide. Although Venturia inaequalis is considered a model species in plant pathology, other Venturia spp. also cause scab on other fruit trees. Relative to the substantial research that has been conducted on V. inaequalis and apple scab, little research has been conducted on Venturia spp. affecting other fruit trees. In this review, the main characteristics of plant-pathogenic species of Venturia are discussed with special attention to V. inaequalis affecting apple, V. pyrina affecting European pear, V. nashicola affecting Asian pear, V. carpophila affecting peach and almond, Fusicladium oleagineum affecting olive, F. effusum affecting pecan, and F. eriobotryae affecting loquat. This review has two main objectives: (i) to identify the main gaps in our knowledge regarding the biology and epidemiology of Venturia spp. affecting fruit trees; and (ii) to identify similarities and differences among these Venturia spp. in order to improve disease management. A thorough review has been conducted of studies regarding the phylogenetic relationships, host ranges, biologies, and epidemiologies of Venturia spp. A multiple correspondence analysis (CA) has also been performed on the main epidemiological components of these Venturia spp. CA separated the Venturia spp. into two main groups, according to their epidemiological behavior: the first group included V. inaequalis, V. pyrina, V. nashicola, and V. carpophila, the second F. oleagineum and F. eriobotryae, with F. effusum having an intermediate position. This review shows that Venturia spp. affecting fruit trees are highly host-specific, and that important gaps in understanding the life cycle exist for some species, including V. pyrina; gaps include pseudothecia formation, ascospore and conidia germination, and mycelial growth. Considering the epidemiological information reviewed, this paper shows that the use of Mills tables to predict infection periods should be avoided for Venturia spp. other than V. inaequalis.
Since the late nineteenth century, apple scab has been extensively investigated, and substantial information-covering all key aspects of the biology and genetics of the fungus and the epidemiology and control of the disease-has been published and reviewed by Machardy (1996) and Bowen et al. (2011). In contrast to the efforts devoted to investigating V. inaequalis, little work has been conducted on Venturia spp. affecting other fruit trees, as shown by the number of papers published for each species ( Figure 1A). This difference in research effort and number of publications, however, does not directly reflect the importance V. pyrina f. sp. piri a Names currently accepted (in bold); for pleomorphic species, names of the anamorphs are indicated after the slash. b Based on Sivanesan (1977) and Schubert et al. (2003).
of the host crops worldwide ( Figure 1B). The difference might be explained by (i) minor investments in these non-apple crops, (ii) less specialized management directed at the non-apple crops, and (iii) the common use of the information developed for V. inaequalis for managing the other fruit scabs. Concerning the last point, researchers generally assume that infection of any scab fungus may occur under environmental conditions similar to those required by V. inaequalis. The Mills and Laplante's (1954) table, which is the most popular system for scheduling fungicides against apple scab, has been broadly recommended for management of pear scab (Sobreiro and Mexia, 2000;Mitcham and Elkins, 2007;Travis et al., 2012;Elkins et al., 2016), cherry scab (Schweizer, 1958), peach scab (Keitt, 1917;Pineau et al., 1991), and loquat scab (Ramos, 2008;GVA, 2014). However, there is no clear evidence that the environmental conditions conducive for infection are similar for all of these Venturia species. In fact, recent studies have revealed important differences concerning the environmental requirements for infection by F. eriobotryae and F. oleagineum vs. V. inaequalis (Viruega et al., 2011;González-Domínguez et al., 2013). In addition, substantial differences exist in the ecophysiologies and the life cycles of their hosts. This review has two main objectives: (i) to identify the main gaps in our knowledge regarding the biology and epidemiology of Venturia spp. affecting fruit trees; and (ii) to identify similarities and differences between these Venturia spp. in order to improve disease management. To accomplish these objectives, the authors have thoroughly reviewed the studies regarding the phylogenetic relationships, host ranges, biologies, and epidemiologies of Venturia spp. Finally, the review discusses the implications of the similarities and differences in the fungi and the diseases for the management of the different scab diseases.
The anamorphs of Venturia spp. have been traditionally classified in three genera: Fusicladium Bonord., Pollaccia Baldacci & Cif., and Spilocaea Fr. Assignment to these genera depends on the morphology of the conidiogenous cells. These cells are sympodial in Fusicladium and percurrent in Pollaccia and Spilocaea (Hughes, 1953;Sivanesan, 1977). Recents works that used both morphological and molecular characters concluded that the anamorphic species of Venturia should not be separated into these three genera because (i) most species have both sympodial and percurrent conidiogenous cells, and (ii) molecular phylogenetic analysis clearly shows that Venturia and its anamorphs are monophyletic group (Schubert et al., 2003;Beck et al., 2005). Because most anamorphs of Venturia have been classified as Fusicladium, this name was proposed to designate the asexual stage of Venturia spp. (Braun et al., 2002), and used in the monograph written by Schubert et al. (2003). Very recently, the International Commission on the Taxonomy of Fungi has proposed the use of Venturia instead of Fusicladium for the species with only anamorph stage, following the guidelines of the "Amsterdam Declaration on Fungal Nomenclature" (May, 2017). In this paper we keep the dual nomenclature because it facilitates to distinguish the species with or without sexual phase.
In 2003, Partridge and Morgan-Jones (2003) proposed the new genus Fusicladosporium, including the anamorphs of pecan (F. effusum), peach (F. carpophillum), and maple scab (F. humile; teleomorph V. acerina). The authors considered that significant morphological differences (conidia formed in chains and the prominence of conidial scars on conidiophores) distinguished these anamorphs from those in the genera Cladosporium and Fusicladium. However, this new taxon seems unjustified because two older generic names for Venturia anamorphs with catenate conidia are available (Hormocladium Höhn. and Ramalia Bat) and because phylogenetic analysis of the ITS region demonstrates that the erection of Fusicladosporium results in a polyphyletic genus (Schubert et al., 2003;Beck et al., 2005). Thus, Fusicladosporium is currently considered a synonym of Fusicladium (Schubert et al., 2003;Crous et al., 2007;Scherm et al., 2008;Seyran et al., 2009;Lalancette et al., 2012).

VENTURIA SPECIES AS FRUIT TREE PATHOGENS Geographical Distribution
Differences exist in the geographical distribution of the Venturia spp. that are pathogens of fruit trees (Figure 2).
There are numerous reports about the worldwide distribution of apple scab (Machardy, 1996;Schnabel et al., 1999;Rossi et al., 2007;Gladieux et al., 2008Gladieux et al., , 2010bXu et al., 2009;Bowen et al., 2011;Li et al., 2011;Padder et al., 2013). Apple scab occurs in every country where apple (Malus x domestica) is cultivated (Machardy, 1996;Gladieux et al., 2008), with the exception of West Australia, where the disease was eradicated (McKirdy et al., 2001). Venturia inaequalis probably emerged in Central Asia, the center of apple origin (Tenzer and Gessler, 1999;Gladieux et al., 2008Gladieux et al., , 2010bXu et al., 2008Xu et al., , 2012, and followed its host's expansion into Europe and, more recently, into regions with the expansion of apple cultivation (Gladieux et al., 2008). Venturia inaequalis infecting apple in Europe and Central Asia consists of three distinct populations: (i) a large European population infecting the domesticated apple and the wild Malus sylvestris; (ii) a large Central Asian population infecting the domesticated apple and populations of Malus sieversii; and (iii) a more geographically restricted population associated with M. sieversii in areas where M. domestica is absent (Gladieux et al., 2010b). Xu et al. (2008Xu et al. ( , 2012 found a higher variability in a population of V. inaequalis from the same orchard in the UK than in populations from different cultivars or regions in China. Overall, V. inaequalis appears to be a model invasive plant pathogen with a broad geographic distribution and well-established populations (Gladieux et al., 2010b).
Like V. inaequalis, V. pyrina has a worldwide distribution that is closely associated with the distribution of its host, the European pear (P. communis; Ogawa and English, 1991; Figure 2). V. nashicola, in contrast, is restricted to China, Japan, South Korea, and Taiwan, where Japanese and Chinese pears are widely cultivated (www.plantwise.org; EPPO, 2016). V. nashicola is currently considered a quarantine organism in the EU, USA, Israel, and Turkey (EPPO, 2016).
Venturia carpophila affecting Prunus spp., F. oleagineum affecting olive, and V. cerasi affecting cherries have relatively restricted distributions, which again reflects the distributions of their hosts (www.plantwise.org) (Figure 2). Venturia cerasi has been reported in Canada, Brazil, New Zealand, Iran, and northern Europe, where cherry is widely cultivated (www.plantwise.org; www.fao.org). Fusicladium oleagineum is widespread in the Mediterranean basin as well as in other temperate and subtropical areas of the world (Graniti, 1993). However, its presence has not been noted in other areas where olives are grown, such as North America, Northern Europe, and South Asia (i.e., India, Nepal, Thailand, and Vietnam; Figure 2). The distribution of V. carpophila is similar to that of F. oleagineum, although the former species has been reported in South Asia but not in South Africa (www.plantwise.org).
Fusicladium effusum and F. eriobotryae apparently have relatively restricted distributions. F. effusum has been detected in South Africa and Australia (Figure 2) and is widely distributed in the Americas, where pecan is an important crop. F. eriobotryae has mainly been reported in the Mediterranean basin, but it was also reported in the USA and Chile (Raabe and Gardner, 1972;Acuña, 2010). As far as known, F. eriobotryae is not present in China, the center of origin of loquat. Moreover, no references were found confirming the presence of F. eriobtryae in Japan or Taiwan, where loquat is widely cultivated.

Phylogenetic Relationships
Several molecular studies have evaluated the phylogenetic relationships among the Venturia species affecting fruit trees (Schnabel et al., 1999;Stehmann et al., 2001;González-Lamothe et al., 2002;Le Cam et al., 2002;Beck et al., 2005;Sánchez-Torres et al., 2009;Gladieux et al., 2010a;Zhao G.-J. et al., 2011;. In these studies, Venturia species and their anamorphs formed a monophyletic clade composed of several small sub-clades. The sub-clades mainly contain strains of the same species, although in some cases strains from different species form a unique sub-clade, as occurs with V. cerasi and V. asperata (Schnabel et al., 1999;Stehmann et al., 2001;González-Lamothe et al., 2002;Beck et al., 2005), V. pyrina and V. nashicola (Schnabel et al., 1999;Stehmann et al., 2001;González-Lamothe et al., 2002;Beck et al., 2005), and V. inaequalis, F. eriobotryae, and F. pyracanthae (Le Cam et al., 2002). All of these analyses concerned the ITS region, which is the most widely used region for species identification in mycology. However, results from this region cannot be considered conclusive for distinguishing closely related fungal species (Kiss, 2012;Schoch et al., 2012). When other nuclear loci have been used, such as the elongation factor or the G3PD genes, differences between closely related Venturia spp. have been observed (Sánchez-Torres et al., 2009;Gladieux et al., 2010b;. V. inaequalis, F. eriobotryae, and F. pyracanthae exhibit a high sequence similarity in their ITS and other DNA regions (Gladieux et al., 2010a). In general, the topology of the phylogram for Venturia species aligns closely with that of the host genera, demonstrating a close co-evolutionary relationship between the pathogenic Venturia spp. and their respective fruit tree hosts (Ishii and Yanase, 2000;Beck et al., 2005;Bowen et al., 2011). Because of these high similarity and based on the criterion of concordance between multiple gene genealogies , Le Cam et al. (2002) and Gladieux et al. (2010a) considered F. eriobotryae and F. pyracanthae as formae speciales of V. inaequalis. Sánchez-Torres et al. (2009) performed further molecular analyses (a phylogenetic analysis of the G3PD gene, a microsatellite-primed PCR analysis, and RAPD fingerprinting) and pathogenicity tests for F. eriobotryae and V. inaequalis; these authors concluded that the loquat scab fungus is a distinct species from V. inaequalis.
Similarly, researchers were able to distinguish between V. pyrina and V. nashicola based on a further phylogenetic analysis of the elongation factor gene , whereas an earlier study was unable to do so based on the ITS region (Beck et al., 2005). Until the 1960s, V. pyrina was considered the causal agent of scab on Japanese and Chinese pear (Pyrus pyrifolia var. culta and P. ussuriensis, respectively). However, further studies indicated that the causal agent of scab on Asian pears differed from V. pyrina (Tanaka and Yamamoto, 1964;Ishii and Yanase, 2000;Abe et al., 2008;.
The results described in the previous paragraphs show that Venturia spp. have a monophyletic evolutionary origin and a close co-evolutionary development with their hosts. In addition, species delimitations become clearer when the phylogenetic analyses are performed in DNA regions in addition to the ITS region.

Host Range
Venturia spp. are confined to six host families: Acaraceae, Betulaceae, Cornaceae, Oleaceae, Rosaceae, and Salicaceae (Sivanesan, 1977). Venturia spp. also seem to be highly host specific in that each species is usually confined to one host genus or at least to allied host genera in one host family (Schubert et al., 2003).
Some controversy exists regarding the host range of the Venturia spp. on fruit trees. In the monograph of Venturia published in 1977, Sivanesan listed 14 species of Rosaceae as hosts of V. inaequalis, and these hosts belonged to the genera Cotoneaster, Malus, Pyracantha, Pyrus, and Sorbus. Schubert et al. (2003) increased the hosts of V. inaequalis to including 12 genera by adding Amelanchier, Aronia, Docynia, Eriobotrya, Heteromeles, Kageneckia, and Prunus. The latter authors, however, did not cite specific studies regarding the ability of V. inaequalis to infect these hosts, and some of these host-pathogen interactions have been refuted (Menon, 1956;Raabe and Gardner, 1972;Ishii and Yanase, 2000;Stehmann et al., 2001;Le Cam et al., 2002;Chevalier et al., 2004;Sánchez-Torres et al., 2007aAbe et al., 2008).
Both Schubert et al. (2003) and Sivanesan (1977) considered P. communis to be a host of V. inaequalis; Schubert et al. (2003) also considered Malus domestica to be a host of V. pyrina. Other studies, however, have failed to obtain infection of pear by V. inaequalis or infection of apple by V. pyrina (Menon, 1956;Stehmann et al., 2001;Chevalier et al., 2004) (Table 2). Menon (1956) inoculated apple and pear plants with three kinds of V. inaequalis and V. pyrina inocula (mycelium plugs, conidial suspensions, and ascospore suspensions) and observed clear scab symptoms on apple only with V. inaequalis and on pear only with V. pyrina and in both cases only with conidia or ascospores ( Table 2). Stehmann et al. (2001) found pseudothecia of both V. inaequalis and V. pyrina in dead apple leaves but cross inoculations with the conidial suspensions of these isolates on detached apple and pear leaves resulted in infection only when apple was inoculated with V. inaequalis and when pear was inoculated with V. pyrina ( Table 2). Although the conidia of V. inaequalis and V. pyrina germinated on both hosts, and although the germ tubes produced appressoria and runner hyphae on both hosts, a dense subcuticular network of stroma was produced only in the compatible host (Stehmann et al., 2001;Chevalier et al., 2004).
Overall, results from cross inoculations of Venturia spp. on different Rosaceous hosts ( Table 2) show that these pathogens are host specific, and the information contained in the monographs of Schubert et al. (2003) and Sivanesan (1977) cannot be reproduced. Host specificity requires further study for F. pyracanthae and F. eriobotryae.
Interestingly, pseudothecia are found in those Venturia spp. affecting deciduous fruit trees such as apple, cherry, nectarine, or European and Asian pears, but not in those species affecting evergreen trees, including pyracantha, loquat, and olive ( Table 1).
The pseudothecia-forming species survive winter mainly as pseudothecia in detached leaves on the surface of the orchard floor. In spring, these pseudothecia repeatedly discharge ascospores, which constitute the primary inoculum for infection. Species that lack pseudothecia overwinter (or oversummer) as mycelium and conidia in lesions on twigs and leaves and in the mummified fruits remaining in the tree after harvest (Figure 3).
The asexual cycle is similar for all Venturia species. Conidia produced on lesions cause secondary infections during the entire tree-growing season as long as the environmental conditions permit conidial production, dispersal, germination, infection, and lesion growth (Figure 3).

Primary Inoculum Sources
Venturia inaequalis survives in winter mainly as pseudothecia on dead, scabbed leaves on the ground ( Table 3; Machardy, 1996). Under some environmental conditions, the fungus can survive in winter as conidia, and when this occurs, these conidia contribute to the primary inoculum. Studies by Holb et al. (2004a) in The Netherlands, Hill (1975) in Germany, and Becker et al. (1992) in New York found that, although large numbers of conidia are present on the surface of shoots or outer bud tissues, conidia are able to overwinter (i.e., are viable in early spring) only on the inner of bud tissues. Becker et al. (1992) and Holb et al. (2004a) reported that this conidia that overwinter in buds are able to infect young green tissues. Recently, Passey et al. (2017) suggested that conidia may contribute 20-50% of the primary inoculum in early spring. Venturia pyrina overwinters as both pseudothecia in affected leaves on the orchard floor and as conidia in twig lesions (Marsh, 1933;Kienholz and Childs, 1937;Williamson and Burchill, 1974;Bearden et al., 1976;Latorre et al., 1985;Spotts et al., 2000;Rossi et al., 2009). Based on conidia and ascospore trappings in England and Oregon, respectively, Marsh (1933) and Kienholz and Childs (1937) considered that primary infections by V. pyrina resulted largely from conidia derived from the previous seasons. Other reports have indicated that ascospores from pseudothecia are the main inoculum source (Bearden et al., 1976;Latorre et al., 1985;Spotts et al., 2000;Rossi et al., 2009) and that infection from conidia overwintering on twig lesions only occur on severely affected pear trees.
Conidia from dormant mycelia in buds of pear trees have been considered the main source of primary inoculum for V. nashicola (Li, 1959;Yin and Yu, 1988). Ascospores of V. nashicola have also been considered an important source of primary inoculum in China (Lian et al., 2006) and Japan (Umemoto, 1990b).
Venturia carpophila overwinters as mycelia in lesions on fruitbearing 1-year-old twigs (Scherm et al., 2008;Lalancette et al., 2012), and the conidia produced on these lesions act as the primary inoculum. Pseudothecia of V. carpophila were found in overwintering apricot leaves only in a severely affected orchard in Australia (Fisher, 1961). However, the epidemiological role of the teleomorph in the V. carpophila disease cycle remains unknown (Lalancette et al., 2012).
On pecan trees, F. effusum overwinters mainly on the surface of twigs and nuts as stromata, which sporulate profusely in early spring (Demaree, 1924). The most important inoculum sources of F. oleagineum are the infected leaves remaining in the olive canopy (Graniti, 1993;Viruega et al., 2013), because the fungus does not produce conidia on fallen, scabbed leaves (Viruega et al., 2013).

Ascospore Discharge and Dispersal
Environmental requirements for ascospore discharge have been frequently studied for V. inaequalis (Table 3). For V. inaequalis, V. pyrina, and V. nashicola, ascospores are mainly discharged during or following rain events (Latorre et al., 1985;Umemoto, 1990a;Spotts and Cervantes, 1994;Machardy, 1996;Villalta et al., 2001;Eguchi and Yamagishi, 2007;Lian et al., 2007;Rossi et al., 2009). Villalta et al. (2001) captured 90% of V. pyrina ascospores during rain events, and Rossi et al. (2009) observed that at least 1.2 mm of rain was neccesary for ascopore ejection. For V. nashicola, Lian et al. (2007) observed that 10 s of wetness was sufficient for ascospore discharge. For V. inaequalis, 0.0025 mm of rain may result in ascospore ejection, but >0.2 mm of rain is usually necessary to capture ascospores in traps (Machardy, 1996). Although some papers have reported the trapping of a few ascospore in periods without rain (Machardy, 1996), the trapping of many ascospores in the early morning following heavy dew at night has been reported only in Norway (Stensvand et al., 1998).
Pseudothecia are able to eject ascospore to a height of 8 mm for V. nashicola (Umemoto, 1990a) and 5-13 mm for V. inaequalis (Aylor and Anagnostakis, 1991). After ejection, ascospores of both pathogens are dispersed by wind. Umemoto (1990a) was able to sample air-borne ascospores of V. nashicola to a distance of 10 m from the inoculum source. In the case of V. inaequalis, ascospores have been captures as far as 45 m from the inoculum source (Holb et al., 2004b).

Ascospore Germination and Infection
Ascospore germination has been studied for V. inaequalis and V. nashicola (Table 3), and had similar temperature requirements for the two species. The ascospores germinate between 5 and 30 • C (germination of V. inaequalis has been observed at 0.5 • C), with the optimum at 15-25 • C (Louw, 1948;Boric, 1985;Lian et al., 2007). Ascospores begin to germinate after 2-3 h when the temperature is optimal; at 10 • C, V. inaequalis germinated after 3 h of wetness and V. nashicola after 6 h of wetness (Boric, 1985;Lian et al., 2007). Keitt and Jones (1926) were the first to conduct a controlledenvironment experiment concerning the minimum number of hours of wetness required for infection by V. inaequalis ascospores. This information was then used by Mills (1944) to develop a chart representing the minimum hours of wetness for light, moderate, and severe infection. Although this publication is considered a milestone in plant pathology, many reports (reviewed by Machardy and Gadoury, 1989) have found that the chart requires modification. Infection by ascospores requires approximately 3 h less than the minimum proposed by Mills, i.e., it requires 5 h of wetness at 20 • C and 8 h of wetness at 12 • or 25 • C. Therefore, Machardy and Gadoury (1989) proposed a new curve that describes the minimum hours of wetness necessary for infection at any temperature. Stensvand et al. (1997) subsequently modified the infection curve for temperatures between 2 and 8 • C. Villalta et al. (2000) reported that infection by V. pyrina ascospores was similar to that reported for V. inaequalis (Machardy and Gadoury, 1989) at temperatures below 10 • C and at 25 • C. In the optimal temperature range of 20-25 • C, however, infection by V. pyrina ascospores required 9 h of wetness rather than the 5 h required by V. inaequalis.

Production of Conidia
Venturia inaequalis, V. nashicola and Fusicladium oleagineum are able to sporulate at temperatures from 5 to 25 • C ( Table 3). V. pyrina can sporulate at temperatures between 5 and 28 • C, which are the only temperatures tested for this species (Ben-Yephet, 1977), whereas V. carpophila can also sporulate at 30 • C (Lalancette et al., 2012). The highest sporulation rate occurred from 15 to 20 • C for all species, except for F. eriobotryae, which produced the most conidia at 5-10 • C (Marras, 1963).

Dispersal of Conidia
Spore sampling studies have been conducted for several species (Table 3). These studies consistently show that conidia of Venturia spp. are mainly collected during or soon after rain events.
Venturia conidiophores (which are short and blunt) do not favor the removal of conidia by wind. Conidia of V. inaequalis were dislodged from dry, sporulating lesions by strong air currents only; the addition of a drop of water, however, caused conidiophores to swell and conidia to dislodge almost Thick lines indicate optimal temperatures. Dotted lines indicate temperatures that are known not to support the process based on experimental evidence. Numbers indicate the hours of wetness necessary at each temperature. For sporulation (A), the RH range in which the process can occur is indicated, with the optimal RH in brackets. For conidial germination (C), the RH range in which the process can occur is indicated; FW indicates that free water is required for germination. Fef, Fusicladium effusum; Fer, Fusicladium eriobotryae; Fol, Fusicladium oleagineum; Vca, Venturia carpophila; Vin, Venturia inaequalis; Vna, Venturia nashicola; Vpy, Venturia pyrina.
immediately (Frey and Keitt, 1925). In several spore-sampling studies in orchards, conidia were trapped from the air, frequently in low numbers, only during rainy weather; in only a few cases were high numbers of V. inaequalis conidia trapped during dry weather (Keitt and Jones, 1926;Gupta and Lele, 1980), with a diurnal periodicity and a peak in the afternoon (Hirst and Stedman, 1961). Machardy (1996) hypothesized that the release of conidia on dry days was triggered by the moisture provided by dew. A steep dispersal gradient was observed for V. inaequalis conidia, with few conidia sampled at >10 m from the inoculum source (Wiesmann, 1932).
For V. pyrina, conidia were also sampled from the air in periods with rain (Kienholz and Childs, 1937). The steep dispersal gradient for V. nashicola conidia (maximum dispersal distance = 8 m) suggests that these conidia are splash dispersed (Umemoto, 1990a).
The role of rain dispersal for V. carpophila was re-evaluated by Lan and Scherm (2003). In a 4-year study, air-borne conidia contributed little to fruit scab in comparison to water-borne conidia; exclusion of splashing conidia decreased disease severity by >90%, and runoff of water from the twig to the fruit via the peduncle also contributed to scab development. Bock et al. (2011) showed that V. carpophila lesions are not uniformly distributed on the peach fruit surface, i.e., most lesions develop near the peduncle.
Conidia of F. oleagineum were mainly dispersed by rain, with a low degree of wind dissemination in the absence of rain when RH was high . Fusicladium oleagineum conidia were dispersed near the inoculum source (<10 m) with a linear and positive relationship between rainfall and numbers of conidia dispersed (Viruega et al., 2013). De Marzo et al. (1993 observed that the psocopteran Ectopsocus briggsi helps spread F. oleagineum conidia by carrying them on its body surface or by allowing them to pass without damage through its alimentary canal. Dispersal of F. eriobotryae conidia was also closely associated with rain. More than 90% of the conidia were collected during rainy periods, and 0.2 mm of rain was sufficient to trigger a dispersal event (González-Domínguez et al., 2014b). A strong aggregation of loquat scab lesions between and within loquat trees also confirmed that F. eriobotryae conidia were mainly splash dispersed (Madden, 1992;González-Domínguez et al., 2014b).

Germination of Conidia
Environmental effects on conidial germination have been studied in vitro for several species (Table 3 and Figure 4). Conidia of all of these species are able to germinate at temperatures between 10 and 30 • C, with the exception of F. oleagineum, whose conidia did not germinate at temperatures >25 • C (Obanor et al., 2007). Germination at 5 • C occurs in all of the species in which it has been tested, i.e., V. inaequalis, V. nashicola, F. effusum, F. oleagineum, and F. eriobotryae (Figure 4). Temperatures >30 • C have been tested only for V. inaequalis and F. effusum, whose conidia were able to germinate at 32 and 40 • C, respectively (Converse, 1956;Boric, 1985). Optimal germination temperatures are close to 20 • C for all of the species, except for V. carpophila and F. effusum, for which the optimum was 25 • C.

Infection by Conidia
The effect of environment on conidial infection has been studied for most of the Venturia spp. considered in this review (Table 3 and Figure 4). For V. carpophila, laboratory experiments have not been performed, and only general requirements were mentioned by Scherm and Brannen (2004). Mills and Laplante (1954) stated that V. inaequalis conidia were able to cause infection in two-thirds of the time required by ascospores.
Subsequent laboratory and field studies have been reviewed by Machardy and Gadoury (1989). For conidial infection, the latter authors used the results of Schwabe (1980) to develop curves for minimum requirements of temperature and wetness duration for infection. The curve developed with data from laboratory studies was similar in shape to the Mill's curve but the time required to infect was greater for the Machardy and Gadoury curves than for the Mills curves.
For all of the species considered in this review, the optimal temperature for infection is 20 • C, but there are differences in the minimum number of hours with high humidity or wetness required for infection (Figure 4). At 20 • C and under continuous wetness, V. nashicola and V. inaequalis are able to infect within 5 and 6 h after inoculation, respectively; V. pyrina and F. effusum require 9 h, and F. oleagineum and F. eriobotryae only cause infection after 12 h of continuous wetness. At 10 • C, the number of hours of continuous wetness required for infection ranged from 10 to 12 h for V. inaequalis, V. pyrina, and V. nashicola, 18 for F. oleagineum, and 24 for F. eriobotryae. F. effusum had similar wetness requirements (>2 h) at 10-35 • C (Gottwald, 1985). The equation of Magarey et al. (2005) showed a similar behavior for V. inaequalis, V. pyrina, V. nashicola, and F. effusum, in that all four species were able to cause infection with only a few hours of wetness under a wide temperature range; F. eriobotryae and F. oleagineum, in contrast, had stricter requirements for both wetness duration and temperature (Figure 5).

Latency Period
Specific experiments on the effect of environmental conditions on the length of latency period (i.e., the time from infection until the occurrence of sporulating lesions) have been conducted only for V. inaequalis and F. oleagineum (Table 3). This period is shorter for V. inaequalis (ranging from 8 days at 18.6 • C to 17 days at 9 • C) than for F. oleagineum (60 days at 6 • C, 16 days at 16 • C and >120 days at 25 • C; Mills, 1946;Roubal et al., 2013). For V. inaequalis, incubation at low RH (60-70%) for >9 days lengthened the latency period and lesions did not develop on plants incubated at low RH (Tomerlin and Jones, 1982). For F. oleagineum, leaf age affects the length of the latency period in laboratory experiments; latency ranged from 31 days in young leaves to 64 days in old leaves (Viruega et al., 2011).  Magarey et al. (2005). For each species, the maximum, minimum, and optimal temperatures for infection, and the minimum hours of wetness required are those indicated in Table 4.  Y  30  1  20  9  32  3  22  10  N  Y   Average  30  3  20  7  31  4  19  11   SD  0  2  0  2  4  2  2  5 CA 1 e X X X a Climatic zones are: C, cold; ST, subtropical; T, tropical. These zones were proposed by Kottek et al. (2006) and are shown in Figure 2. b Y and N indicate that the host is or is not deciduous. c TM, maximum temperature for infection; Tm, minimm temperature for infection; To, optimal temperature for infection; Wm, minimum hours of wetness for infection. d Y indicates that the species can germinate only in free water; N indicates that the species can germinate at <100% RH. e For CA1, CA2, and CA3, X indicates the components used in each analysis.

MULTIVARIATE ANALYSIS OF THE EPIDEMIOLOGICAL COMPONENTS
As part of the current review, a multiple correspondence analysis (CA) was performed on epidemiological components of seven Venturia spp. CA is a multivariate statistical method that makes it possible to represent contingency tables in a pictorial form (Savary et al., 1995). CA is based on a raw data matrix, in which the rows are the objects and the columns are the variables. In this case, the objects are the seven Venturia spp.-V. carpophila, V. inaequalis, V. pyrina, V. nashicola, F. effusum, F. eriobotryae, and F. oleagineum-for which epidemiological information exists for seven qualitative variables and eight quantitative variables ( Table 4).
In our case, qualitative variables (e.g., the known presence of the sexual stage in nature) were classified using Yes or No; quantitative variables (e.g., optimal temperature for conidial infection) were ranked as high, medium, or low based on the average ± standard deviation (SD) of each data set. For example, the maximum temperature for conidial infection (TM) of the seven Venturia spp. ranged from 25 to 37 • C, with an average of 31 • C and a SD of 4. Thus, Venturia spp. in which the maximum temperature for infection was TM ≤ 27 • C (i.e., 31-4) were classified as low, those with 27 > TM < 35 • C were classified as medium, and those with TM ≥ 35 • C were classified as high ( Table 4).
The data matrix of Table 4 was used to perform CA with two dimensions (D1 and D2) using the multiple correspondence analysis procedure of SPSS (ver. 23; SPSS Inc.). Three analyses were performed: (i) with all components (CA1); (ii) with all components except those concerning the distribution in different climate types (CA2); and (iii) with only those components concerning the asexual stage (CA3). These analyses accounted for 80.0, 88.5, and 98.2% of data variance, respectively (Table 5).
Overall, these analyses separated the Venturia spp. into two main groups when the seven species were plotted on the D1-D2 space (Figure 6). The first group contained V. inaequalis, V. pyrina, V. nashicola, and V. carpophila, and the second group contained F. oleagineum and F. eriobotryae, with F. effusum having an intermediate position depending on the CA analysis.
In CA1 (Figure 6A), grouping of the species was mainly based on the presence/absence of the sexual stage, infection of deciduous vs. non-deciduous trees, and the thermal and moisture requirements for ascosporic infection. These variables had the most influence (i.e., had high discrimination values, DVs) in D1, which accounted for 53.3% of the variance in the dataset ( Table 5); the DV was >0.829 for these variables ( Table 6). In CA1, V. nashicola was far from the other species in D2 (Figure 6A), mainly because of its different distribution among climate types (presence/absence of the species in tropical and subtropical climates had DVs = 0.514 and 0.637, respectively; Table 6); the presence/absence of the species in tropical and subtropical climates together with wind dispersal of conidia were the most influential variables in D2 ( Table 6). This result may be biased by the distribution of V. nashicola being restricted to China, Japan, South Korea, and Taiwan, a restriction that may be caused more by quarantine measures by other countries than by differences in environmental requirements. This inference is supported by CA2, in which the pathogen distribution among climates was not considered. In CA2 (Figure 6B), the species were grouped mainly based on the presence/absence of the sexual stage and environmental requirements for ascosporic infection in D1 (Table 6), which accounted for 63.6% of the variance ( Table 5). In CA3 (Figure 6C), the grouping was mainly determined by moisture conditions for conidial germination, infection, and dispersal in D1 (Table 6), which accounted for 60.9% of the variance ( Table 5). The conidia of V. inaequalis, V. pyrina, V. nashicola, and V. carphophila are similar in that they require only a few hours of wetness (6-12 h) to infect and are capable of being dispersed by wind if dislodged by rain or dew. Fusicladium oleagineum and F. eriobotryae require longer periods of wetness to infect (>18 h), have lower maximum temperatures for infection (25-28 • C), and have conidia that germinate only in the presence of free water. In CA3, F. oleagineum and F. eriobotryae had similar scores in D1, but they were far apart in D2 ( Figure 6C) because of different temperature requirements for conidial infection (Tables 4, 6). Both F. effusum and the Venturia group can infect with <12 h of wetness and have wind-dispersed conidia that do not require free water to germinate; however, the temperature requirements for conidial infection differ between F. effusum and the Venturia group.
The presence of two main groups of species probably reflects pathogen adaptation to host ecophysiology. This hypothesis is supported by the monophyly of the genus Venturia (Ishii and Yanase, 2000;Beck et al., 2005;Gladieux et al., 2010a;Bowen et al., 2011). F. eriobotryae and F. oleagineum are both pathogens of Mediterranean plants (loquat and olive, respectively). They are adapted to a warm and dry climate in which the low annual rainfall is distributed mainly in autumn and spring (Csa climate class; Graniti, 1990;Kottek et al., 2006). The absence of the teleomorph in nature may be related to the mild winter temperatures, which can be survived without a quiescent stage. In these fungi, dispersal of conidia occurs only during rain events, perhaps because conidia dispersed in water have a higher probability of germinating and causing infection. The requirement of free water for conidial germination and long periods of wetness for conidial infection may also be adaptations to a dry climate. These requirements would prevent the initiation of an infection cycle in the driest periods of the year. F. eriobotryae and F. oleagineum also have low mycelium growth rates and long latent periods. The evergreen habitus of their hosts means that these species do not require short infection cycles, because the trees are susceptible throughout the year and because the  Table 4); (B) Joint plot of the correspondence analysis performed with 12 components related to epidemiological variables of the sexual and asexual phase (CA2 in Table 4); (C) Joint plot of the correspondence analysis performed with six components related to epidemiological variables of the asexual phase (CA3 in Table 4).  Table 4. b Climatic zones were proposed by Kottek et al. (2006) and are shown in Figure 2. c Whether or not the host is deciduous. d TM, Maximum temperature for infection; Tm, minimum temperature for infection; To, optimal temperature for infection; Wm, minimum hours of wetness for infection. e Whether conidia can germinate only in free water. f Indicates that this epidemiological component was not included in the analyses, as indicated in Table 4.
inoculum may survive on the tree during the season in which environmental conditions are not suitable for sporulation and infection.
Venturia spp. that attack deciduous trees, in contrast, require a sexual stage to survive the winter, when there is no host tissue to be infected and temperatures are low. Their conidia can germinate in the absence of free water, and infection requires fewer hours of wetness, especially in the case of V. inaequalis and V. nashicola. Together, these factors result in relatively short infection cycles, with a higher probability of occurrence. In this case, the deciduous habitus of the host makes the occurrence of infection obligatory, because the main inoculum source for the next season will be the fallen, scabbed leaves, even though overwintering in twig lesions and/or buds is possible.

CONCLUSIONS AND IMPLICATIONS FOR SCAB MANAGEMENT
This review has considered several important aspects of the phylogeny, host range, and life cycle of Venturia spp. affecting fruit trees. These species are responsible for some of the most important diseases of their hosts. Venturia spp. affecting fruit trees are highly host-specific, as indicated by the general failure to obtain infection by crossinoculation (Menon, 1956;Raabe and Gardner, 1972;Ishii and Yanase, 2000;Stehmann et al., 2001;Le Cam et al., 2002;Chevalier et al., 2004;Sánchez-Torres et al., 2007aAbe et al., 2008). This conclusion partially contradicts some previous reviews of the genus Venturia (Sivanesan, 1977) and Fusicladium (Schubert et al., 2003), and has implications for the management of scab diseases in areas where different host species are grown, as is the case, for instance, in the Emilia-Romagna region of North Italy where apple, pear, cherries, and peaches are grown in close proximity. The inoculum produced in one crop cannot infect another crop, and management of a species of Venturia in one orchard has no effect on neighboring crops of other host species.
Although Venturia is one of the most famous and studied genera of plant pathogens, important gaps in understanding the life cycle still remain for some species. This is particularly unexpected for V. pyrina because of the worldwide importance of the crop and because the gaps involve key aspects of the life cycle, including pseudothecia formation, ascospore and conidia germination, and mycelial growth. The only work regarding dispersal of V. pyrina conidia was published in 1937 (Kienholz and Childs, 1937). In most cases, researchers have assumed that V. pyrina requirements are similar to those of V. inaequalis. This assumption, however, is not valid for ascosporic infection because ascospores of V. pyrina require more wet hours to infect (Figure 4). In the case of V. carpophila, specific experiments on the requirements for infection have never been performed, and this limits our ability to correctly manage the disease.
Venturia spp. can infect several parts of the host trees, but the main damage usually results from fruit infection, except for F. oleagineum affecting olive leaves, which can result in important economic losses. For this reason, fruit growers are generally riskadverse to scab diseases and schedule a high number of fungicide applications to achieve a very high level of disease control. Machardy (1996) reported that USA apple growers schedule fungicide applications to achieve <1% of scabbed fruit at harvest. Similar thresholds are usually assumed for loquat, a high value fruit crop in Europe (E. Soler, personal communication). Current trends in disease management aim to avoid this high number of treatments, which involve risks to human health and the environment, and encourage the use of decision support systems (DSSs) (Rossi et al., 2012).
The Mills tables represent one of the first and better known DSSs and have been widely accepted by growers and advisors (Machardy, 1996). However, the use of these tables should be avoided for Venturia spp. other than V. inaequalis. For V. pyrina, F. oleagineum, F. effusum, and F. eriobotryae the Mills tables over-predict the number of scab infections because temperature and wetness requirements of these fungi are different from those of V. inaequalis (Figure 6; Gottwald, 1985;Villalta et al., 2000;Viruega et al., 2011;González-Domínguez et al., 2013). Moreover, the Mills tables can over-predict the number of infections even for apple scab (Machardy and Gadoury, 1989).
For some of the diseases considered in this review, epidemiological models have been developed to predict disease development ( Table 7). Most of these models are simple and consider only one component of the pathogen life cycle, mainly ascospore maturation or infection. A main constraint of these models is that they have never been validated against independent data, i.e., model output (the prediction) has not been compared with a data set of real-world observations different from that used for model development (Rossi et al., 2010). Before these models are used in practical disease control, a robust validation with real data should be performed; the validation data should be obtained from different areas with different epidemiological conditions and for several years (Rossi et al., 2010).
In the case of V. inaequalis, mechanistic weather driven models have been developed for primary infections  and for the whole life cycle (Xu et al., 1995). A mechanistic approach has also been used to develop epidemiological models for V. nashicola and F. eriobotryae, and these models consider most of the components of the life cycle González-Domínguez et al., 2014a). Several advantages have been previously reported for mechanistic vs. empirical models (Caffi et al., 2007;De Wolf and Isard, 2007;Rossi et al., 2010). Mechanistic models attempt to capture the full complexity of the pathogen life cycle and are generally considered to have greater explanatory ability than purely empirical models (De Wolf and Isard, 2007). A main advantage of the mechanistic models is that they can easily incorporate information from previous experiments regarding pathogen biology and epidemiology. Thus, this review should help researchers develop mechanistic models for those scab diseases that currently lack such models. Until such mechanistic models are available, however, the Magarey curves, developed in this review and which identify the requirements for infection, could be used as a starting point to predict infection risk, especially for V. pyrina and F. effusum.

AUTHOR CONTRIBUTIONS
All the authors contributed to the writing of the manuscript. EG and VR performed the data analyses.