Morphology, phylogeny and pathogenicity of Botryosphaeria and Neofusicoccum species associated with drupe rot of olives in southern Italy

Authors


*E-mail: alp@fct.unl.pt

Abstract

Species of Botryosphaeria and Neofusicoccum are well known as pathogens of woody hosts. In this study the species that occur on rotting olive drupes in the main production areas of southern Italy were studied. Species were identified from the morphology of their conidial states in culture and from sequence data of the ITS rDNA operon and partial sequence of the translation elongation factor 1-α gene. Botryosphaeria and Neofusicoccum species were isolated from more than 60% of the affected drupes, suggesting that they are the main contributors to this disease. The most common species was B. dothidea, which was isolated from 34% of the drupes. However, N. australe and N. vitifusiforme were also common and were isolated from 16 and 12%, respectively. Two other species (N. parvum and N. mediterraneum) were uncommon and occurred on less than 1% of the drupes. All five species were pathogenic on the two cultivars of olive tested. The most aggressive species was N. vitifusiforme, followed by N. australe and B. dothidea. The two olive cultivars differed in their susceptibility to the pathogens. The results show that B. dothidea, N. vitifusiforme and N. australe are important pathogens of olives.

Introduction

The European olive (Olea europaea) is an important crop in southern Italy. According to the Italian National Institute of Statistics (http://www.istat.it) olives were cultivated on 906 k ha throughout this region in 2006. Of the 2960 k tonnes of olives harvested, 2908 k tonnes were destined for oil production while the remainder were used as table olives. The greatest production was in the Calabria (1111 k tonnes) and Puglia (1090 k tonnes) regions with significant production in Sicilia (286 k tonnes) and Basilicata (42 k tonnes). Thus, the olive industry contributes significantly to the economy and employment of the southern regions of Italy.

Diseases of olive drupes can cause financial losses through direct loss of rotted drupes, reduced cosmetic value of table olives and reduced quality of the oil due to fungal infections. González et al. (2006) showed that oil from infected drupes had higher acidity, higher peroxide levels and lower stability than oil from healthy drupes.

The disease of olive drupes known locally as ‘Lebbra’ is caused by Colletotrichum acutatum. This disease is easily recognized by the large quantities of orange conidia that are produced by the pathogen. A similar disease, but distinctive in the absence of orange conidial masses, has been known for many years and appears to be widespread throughout the Mediterranean region (Trapero & Blanco, 2004). This disease is commonly known as ‘drupe rot’ but it has also been referred to as ‘Dalmation disease’ (González et al., 2006).

Since it was first described in 1883, the causal agent of drupe rot has undergone several taxonomic changes. It was first described as Phyllosticta dalmatica and subsequently transferred to other genera such as Phoma dalmatica, Macrophoma dalmatica, Sphaeropsis dalmatica and Camarosporium dalmaticum. Phillips et al. (2005a) determined that all these names are synonyms of Fusicoccum aesculi, which is the anamorph of Botryosphaeria dothidea. They further determined that this fungus is the main cause of drupe rot of olives in central Greece.

A recent phylogenetic study of the Botryosphaeriaceae revealed that Botryosphaeria is composed of several distinct lineages that correspond to individual genera (Crous et al., 2006). Only B. dothidea and B. corticis were retained in Botryosphaeria, while other species with Fusicoccum-like anamorphs were transferred to the new genus Neofusicoccum.

The only Neofusicoccum species that has been reported on olives is Neofusicoccum ribis (= Botryosphaeria ribis) (Romero et al., 2005). Nevertheless, Neofusicoccum spp. are known to be pathogens of a wide range of woody hosts causing diebacks and fruit rots (Pennycook & Samuels, 1985; Phillips et al., 2002; Denman et al., 2003; Niekerk et al., 2004; Slippers et al., 2005). There have been no detailed studies of the Botryosphaeria and Neofusicoccum species that occur on olives.

In a recent survey of drupe rot in southern Italy, a large collection of isolates corresponding morphologically to Fusicoccum and Neofusicoccum species were isolated. The aim of the present study was to determine the identity of the species and to determine their pathogenicity on two olive cultivars. Species were identified from their morphology and a study of combined ITS and EF1-α sequence data.

Materials and methods

Isolates and morphology

Between 70 and 100 olive drupes with symptoms of drupe rot were collected from each of 23 localities in the main olive producing regions of southern Italy (Table 1). The drupes were examined with a dissecting microscope and isolations were made by spreading the cirrus of conidia exuding from pycnidia on potato dextrose agar (PDA) (Difco) supplemented with 500 µg mL−1 streptomycin (PDAS). After incubating at 25°C for 18 h, single germinating conidia were transferred to fresh PDA. When no cirri could be seen, the drupes were washed with running tap water for 10 min, dipped in 70% ethanol for 2 min, and pieces of tissue taken from the edge of the lesion and transferred to PDAS. Putative Botryosphaeria and Neofusicoccum isolates, recognized by their rapidly growing colonies with grey mycelium, were transferred to fresh plates of PDA. Identity of the two genera was confirmed from conidial morphology and mode of conidiogenesis (Crous et al., 2006). Isolates were stored on PDA slopes at 5°C. Representative cultures were deposited at the Centraalbureau voor Schimmelcultures (CBS), Utrecht, the Netherlands.

Table 1.  Isolates of Botryosphaeria and Neofusicoccum species from olive drupes in southern Italy used in this study Thumbnail image of

For studies on growth rates and colony morphology, isolates were grown on half strength-PDA and incubated at 25°C in darkness (for growth rates) or on the laboratory bench where they received diffused daylight (for colony morphology). To induce sporulation, cultures were grown on 2% w/v water agar bearing pieces of sterilized pine needles and kept on the laboratory bench. Conidia oozing from the pycnidia were transferred to a drop of water on a glass slide and when the water had almost completely dried out, a drop of 100% v/v lactic acid was added and a coverslip applied. The conidiogenous layer was dissected out from the pycnidia and mounted in 100% lactic acid. Microscope images were recorded with a Leica DFC320 digital camera from images recorded with the ×100 objective. Conidia were measured with the Leica IM500 measurement module. The mean, standard deviation and 95% confidence limits were calculated from at least 20 conidia of each isolate.

DNA isolation and amplification

DNA was isolated from fungal mycelium by the method of Möller et al. (1992). PCR reactions were carried out with Taq polymerase, nucleotides and buffers supplied by MBI Fermentas and PCR reaction mixtures were prepared according to Alves et al. (2004), with the addition of 5% v/v DMSO to improve the amplification of some difficult DNA templates. All primers were synthesized by MWG Biotech AG. The ITS region was amplified using the primers ITS1 and ITS4 (White et al., 1990) as described by Alves et al. (2004). The primers EF1-728F and EF1-986R (Carbone & Kohn, 1999) were used to amplify part of the EF1-α gene as described by Phillips et al. (2005b).

The amplified PCR fragments were purified with the JETQUICK PCR Purification Spin Kit (GENOMED). Both strands of the PCR products were sequenced by STAB Vida Lda. The nucleotide sequences were read and edited with FinchTV 1·4·0 (Geospiza Inc.; http://www.geospiza.com/finchtv). All sequences were checked manually and nucleotide arrangements at ambiguous positions were clarified using both primer direction sequences. Sequences were deposited in GenBank (Table 1).

Phylogenetic analyses

Sequences of the strains isolated in this study were assembled with nucleotide sequences of additional isolates of Botryosphaeria and Neofusicoccum spp. retrieved from GenBank (Table 2). The sequences were aligned with ClustalX version 1·83 (Thompson et al., 1997) using the following parameters: pairwise alignment gap opening = 10, gap extension = 0·1; multiple alignment gap-opening = 10, gap extension = 0·2; delay divergent sequences = 25%; transition weight = 0·5. Alignments were checked and manual adjustments were made where necessary. Phylogenetic analyses of sequence data were done using PAUP* v.4·0b10 (Swofford, 2003). The trees were rooted to Diplodia seriata and D. mutila and visualized with TreeView (Page, 1996).

Table 2.  Additional isolates included in the phylogenetic study of species associated with drupe rot of olives
Isolate numberaSpeciesCollectorHostLocalityGenBank Accession No.
ITSEF
  • a

    Designation of isolates and culture collections: CAA = A. Alves, Universidade de Aveiro, Portugal; CAP = AJL Phillips, Universidade Nova de Lisboa, Portugal; CBS = Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands; CMW = MJ Wingfield, FABI, University of Pretoria, South Africa; STE-U = University of Stellenbosch, South Africa.

CMW 9072Neofusicoccum australeJ. RouxAcacia sp.AustraliaAY339260AY39268
CBS 119046N. australeE. DiogoRubus sp.PortugalDQ299244EU017541
STE-U 4425N. australeF. HalleenVitis viniferaSouth AfricaAY343388AY343347
CBS110299Neofusicoccum luteumA.J.L. PhillipsVitis viniferaPortugalAY259091AY573217
CMW 9076N. luteumS.R. PennycookMalus x domesticaNew ZealandAY339257AY339265
CBS 110887Neofusicoccum vitifusiformeJ. Van NiekerkVitis viniferaSouth AfricaAY343383AY343343
CBS 110880N. vitifusiformeJ. Van NiekerkVitis viniferaSouth AfricaAY343382AY343344
WAC 12398Dichomera eucalyptiT. Burgess/K.-LGoeiEucalyptus diversicolorAustraliaAY744371DQ093214
CMW 15952D. eucalyptiT. Burgess/K.-LGoeiEucalyptus diversicolorAustraliaDQ093194DQ093215
CMW 15953D. eucalyptiT. Burgess/K.-LGoeiEucalyptus diversicolorAustraliaDQ093195DQ093216
CAA 002Neofusicoccum mediterraneumT.J. MichailidesPistacia vera var. ‘Kerman’USAEU017537EU017538
CBS 112878Neofusicoccum viticlavatumF. HalleenVitis viniferaSouth AfricaAY343381AY343342
CBS 112977N. viticlavatumF. HalleenVitis viniferaSouth AfricaAY343380AY343341
CBS 110301Neofusicoccum parvumA.J.L. PhillipsVitis viniferaPortugalAY259098AY573221
CMW 9081N. parvumG.J. SamuelsPopulus nigraNew ZealandAY236943AY236888
CBS 121·26Neofusicoccum ribisN.E. StevensRibes rubrumUSAAF241177AY236879
CBS 115475N. ribisB. SlippersRibes sp.USAAY236935AY236877
CBS 118531Neofusicoccum mangiferumG.I. JohnsonMangifera indicaAustraliaAY615185DQ093221
CBS 118532N. mangiferumG.I. JohnsonMangifera indicaAustraliaAY615186DQ093220
CBS 115791Neofusicoccum eucalyptorumH. SmithEucalyptus grandisSouth AfricaAF283686AY236891
CMW 10126N. eucalyptorumH. SmithEucalyptus grandisSouth AfricaAF283687AY236892
CBS 115766Neofusicoccum eucalypticolaM.J. WingfieldEucalyptus rossiiAustraliaAY615143AY615135
CMW 6539N. eucalypticolaM.J. WingfieldEucalyptus rossiiAustraliaAY615141AY615133
CBS 116741Botryosphaeria dothideaI. RumbosOlea europaeaGreeceAY640254AY640257
CBS 110300B. dothideaA.J.L. PhillipsPopulus nigraPortugalAY640253AY640256
CBS 110302B. dothideaA.J.L. PhillipsVitis viniferaPortugalAY259092AY573218
CBS 115476B. dothideaB. SlippersPrunus sp.SwitzerlandAY236949AY236898
CBS 119048Botryosphaeria corticisP.V. OudemansVaccinium corymbosumUSADQ299246EU017540
CBS 119047B. corticisP.V. OudemansVaccinium corymbosumUSADQ299245EU017539
CBS 112553Diplodia mutilaA.J.L. PhillipsVitis viniferaPortugalAY259093AY573219
CBS 112555Diplodia seriataA.J.L. PhillipsVitis viniferaPortugalAY259094AY573220

The HKY85 nucleotide substitution model (Hasegawa et al., 1985) was used for distance analysis. All characters were unordered and of equal weight. Bootstrap values were obtained from 1000 neighbour joining (NJ) bootstrap replicates.

Maximum parsimony (MP) analyses were performed using the heuristic search option with 1000 random taxa additions and tree bisection and reconnection (TBR) as the tree swapping algorithm. All characters were unordered and of equal weight and gaps were treated as a fifth character. Branches of zero length were collapsed and all multiple, equally parsimonious trees were saved. The robustness of the most parsimonious trees was evaluated by 1000 bootstrap replications (Hillis & Bull, 1993). Other measures used were consistency index (CI), retention index (RI), and homoplasy index (HI). A partition homogeneity test was done with PAUP to asses the validity of combining the ITS and EF1-α data.

Pathogenicity

Pathogenicity of the 41 selected isolates was tested on healthy drupes of the olive cultivars Coratina and Ogliarola. The isolates were grown on PDA at 25°C for 7 days prior to inoculation.

Olive drupes were collected from trees in the Molfetta area in the province of Bari, Puglia in January 2007. Five drupes for each combination of cultivar and isolate were washed with 1% v/v Tween 20 and placed on sterilized metal grids inside 20 cm diameter glass Petri dishes lined with filter papers, that were kept moist throughout the experiment. Drupes were damaged at six closely spaced points by pricking with a sterile needle (area wounded = 16 mm2). A small piece of agar (4 × 4 mm) colonized with the test strain of each isolate was placed on the wounded zone. Controls were inoculated with pieces of uncolonized PDA. The Petri dishes were sealed with Parafilm and incubated at 22 ± 2°C in darkness for 20 days. There were five replicates for each combination of isolate and cultivar arranged as a fully randomized experiment. Aggressiveness of isolates was determined after 20 days from the percentage of the surface showing signs of rotting, which was assessed by referring to a series of standard photographs of drupes with various degrees of rotting. Significance of differences in susceptibility of the two cultivars was gauged by a t-test, while differences in aggressiveness of isolates was determined by analysis of variance.

Results

Isolations

Between October 2000 and December 2005, 1747 rotted drupes were collected from 23 sites in southern Italy. Isolations on PDAS resulted in 1092 isolates of Botryosphaeria and Neofusicoccum spp. (Table 1). The other 655 rotted olives yielded species of Fusarium, Colletotrichum, Alternaria and Phoma, which were not considered further in this study. The Botryosphaeria and Neofusicoccum isolates from each locality were grouped according to their morphology. This resulted in 41 groups of isolates, and one isolate from each group was selected for detailed studies on their morphology, for pathogenicity testing and for sequence analysis (Table 1).

Phylogenetic analysis

ITS phylogeny

PCR products of ~580 bp were obtained for all isolates and ~520 bp were used in the phylogenetic analysis. After alignment the dataset, composed of 41 isolates from olives and 34 sequences retrieved from GenBank, consisted of 537 characters including gaps. Of these, 412 were constant and 12 were parsimony uninformative. Maximum parsimony analysis of the remaining 113 parsimony-informative characters resulted in four trees of 174 steps (TreeBase Accession No. SN 3487). The four trees differed only in the arrangement of isolates in the terminal clades, while the overall topology was the same.

Two major clades were resolved in the MP analysis. One clade corresponded to Neofusicoccum species while the other consisted of two species of Botryosphaeria, namely B. dothidea and B. corticis. Most of the isolates from olives (23) clustered in the B. dothidea clade. The remaining 18 olive isolates fell within four sub-clades in the Neofusicoccum clade. Nine isolates clustered close to N. australe and N. luteum and seven isolates clustered with N. vitifusiforme and Dichomera eucalypti. A single isolate clustered with N. parvum and another with some isolates from pistachio. However, species in the Neofusicoccum clade were not clearly resolved and bootstrap support for the branches was generally low. Therefore, representative isolates from the Botryosphaeria clade and the four groups within Neofusicoccum were selected for sequencing of the EF1-α gene (Table 1).

Combined ITS and EF phylogeny

A partition homogeneity test showed no significant difference (P = 0·27) between the data from the different gene regions, indicating that they could be combined in a single dataset. The sequence alignment of 20 isolates from olives, 30 sequences from GenBank and two outgroup species consisted of 536 characters for the ITS region and 307 for the EF1-α gene, including alignment gaps. The combined dataset consisted of 843 characters, of which 522 were constant and 39 were parsimony-uninformative. Maximum parsimony analysis of the remaining 282 parsimony-informative characters resulted in seven most parsimonious trees (TL = 509 steps, CI = 0·845, RI = 0·964, HI = 0·155). The seven trees differed only in the arrangement of isolates within the terminal clades while their overall topology was the same. One of the trees is shown in Fig. 1. Neighbour-joining analysis resulted in a tree with essentially the same topology as the MP tree.

Figure 1.

One of seven most parsimonious trees generated from combined ITS and EF1-α sequence data of isolates of Botryosphaeria, Dichomera and Neofusicoccum species. Isolate names in bold type are from olives. Maximum parsimony bootstrap values from 1000 replicates are given above the nodes, with neighbour joining bootstrap values below the nodes. The tree was rooted to Diplodia seriata and D. mutila.

The combination of ITS and EF1-α revealed 10 well-supported clades in Neofusicoccum and two in Botryosphaeria. The isolates from olives clustered within four Neofusicoccum clades and with B. dothidea. The combined dataset confirmed that a single isolate belonged to N. parvum and another single isolate clustered with isolates of N. mediterraneum from pistachio. The remaining Neofusicoccum isolates clustered with either N. australe (nine isolates) or in a clade composed of isolates of Dichomera eucalypti and N. vitifusiforme (seven isolates). All the olive isolates that clustered with N. australe lay within a single sub-clade of this species, with low MP bootstrap support (64%) but high NJ bootstrap support (99%). Despite the high NJ bootstrap support, only a single base pair difference in ITS and 3 bp in EF 1-α separated the olive isolates from N. australe.

Morphology

Most of the isolates sporulated within 14 days of incubation on pine needles on water agar. Morphology of the isolates within the B. dothidea clade were typical of that species in their cylindrical conidiogenous cells (Fig. 2a,b) and fusiform, hyaline, aseptate conidia (Fig. 2c) that measured (11·7–)21·4–22·4(–33·3) × (3·9–)5·8–6(–8·5) µm. In some isolates, the conidium wall became darker and thicker, and some conidia developed one or two septa (Fig. 2d,e).

Figure 2.

Botryosphaeria and Neofusicoccum species isolated from olive drupes. (a–e) B. dothidea. (a) conidiogenous layer; (b) conidiogenous cells and developing conidia; (c) hyaline conidia; (d–e) darkened, thicker walled, occasionally septate conidia. (f–g) N. australe. (f) conidiogenous cells; (g) conidia. (h–j) N. vitifusiforme. (h–i) conidiogenous cells; (j) conidia. (k–l) N. parvum. (k) conidiogenous cells; (l) conidia. (m–n) Neofusicoccum mediterraneum. (m) conidiogenous cells; (n) conidia. Bars = 10 µm.

Conidia of isolates in the N. australe clade were hyaline, aseptate and sub-fusiform (Fig. 2f,g) measuring (16·2–)18·4–19·5(–20·7) × (5·8–)6·4–6·8(–7·6) µm, which is somewhat smaller than is typical for N. australe but similar to the dimensions reported for N. luteum. In culture, these isolates produced a yellow pigment that after 5–6 days became violaceous and then darkened, which is typical for both N. australe and N. luteum. Conidiogenous cells of the isolates in the N. vitifusiforme/D. eucalypti clade were long and cylindrical (Fig. 2h,i) and the conidia (Fig. 2j) were fusiform measuring (17–) 18·5–22·5(–22·8) × (4·5–)5–6(–6·5) µm, which is within the range reported for that species. Morphologically the single isolate of N. parvum (Fig. 2k,l) corresponded well with the description by Pennycook & Samuels (1985). A single isolate of N. mediterraneum (CAP253) grouped with isolates from pistachio (Fig. 2m,n). The Dichomera synanamorph was not seen in any of the Neofusicoccum species studied here.

Pathogenicity

A preliminary analysis of the data indicated that no significant differences could be detected between isolates of each species. Therefore the data for all isolates of a given species were pooled (Table 3) so that differences in aggressiveness of the species could be tested. A Student's t-test on the pooled data for each cultivar indicated significant differences in the severity of disease (t0·05 = 5·33, P < 0·001) with cv. Coratina being the most susceptible. Analysis of variance indicated significant differences in the aggressiveness of the five species on cv. Coratina (F0·05 4, 200 = 5·559, P < 0·001) and on cv. Ogliarola (F0·05 4, 200 = 6·438, P < 0·001). On both cultivars the most aggressive species was N. vitifusiforme followed by N. australe and B. dothidea. The data for N. parvum and N. mediterraneum are less reliable because only a single isolate of each was available and the degree of the variability within each species cannot be judged.

Table 3.  Percentage disease severity on olive drupes inoculated with different fungi as determined from the percentage of each drupe rotted
SpeciesNumber of isolatesCultivar
CoratinaOgliarola
Botryosphaeria dothidea2336·615·9
Neofusicoccum australe 951·124·4
Neofusicoccum vitifusiforme 773·434·6
Neofusicoccum parvum 120·020·0
Neofusicoccum mediterraneum 180·0100
Means (n = 5) 46·7323·12
LSD 5% for species comparisons ±6·45±5·56

Discussion

This study represents the first attempt to characterize the species of Botryosphaeria and Neofusicoccum associated with drupe rot of olives in an extensive collection of isolates, and integrating morphology, pathology and molecular data. Preliminary identifications were based on morphology of the anamorphs in culture. Although ITS sequence data allowed definitive identification of B. dothidea it was necessary to include sequence data from the translation elongation factor 1-α to identify unambiguously the Neofusicoccum species. In the present study, five species were identified on olives, of which only one (B. dothidea) has previously been associated with this host (Phillips et al., 2005a; González et al., 2006).

The five species isolated from olive drupes were identified from a combination of morphological (conidial) and molecular (ITS and EF1-α sequence) characters. The B. dothidea isolates clustered with an ex-epitype strain of this species in both the ITS tree and the tree constructed from combined ITS + EF1-α sequence data. The strains of N. vitifusiforme isolated from olives clustered with isolates of N. vitifusiforme from South African grapevines and isolates of Dichomera eucalypti from Australia. Only 2 bp in ITS separated the isolates in this clade, while EF1-α sequences for all these isolates were identical. Barber et al. (2005) showed that several ‘Botryosphaeria’ species from Eucalyptus form a Dichomera synanamorph in culture, but they did not associate it with any Neofusicoccum species. When Crous et al. (2006) introduced Neofusicoccum they stated that the formation of a Dichomera synanamorph is the main feature that differentiates it from Fusicoccum. Since the isolates studied by Barber et al. (2005) clustered with ex-type cultures of N. vitifusiforme in this study, it is clear that D. eucalypti is the synanamorph of N. vitifusiforme. If this is accepted, the host range of N. vitifusiforme can be broadened to include Eucalyptus and Olea.

Of the 41 strains studied, nine clustered as a sub-clade within the N. australe clade. Although this sub-clade was supported by a high NJ bootstrap value (99%), MP bootstrap was low (64%) and only 1 bp difference in ITS and 3 bp differences in EF1-α separated these isolates from N. australe. All these isolates produced a yellow pigment typical of N. luteum and N. australe. Phylogenetically they were considered to be N. australe, but morphologically they were closer to N. luteum. Until more isolates from a wider geographic range have been studied, it is preferable to consider these isolates as N. australe.

Botryosphaeria and Neofusicoccum species were isolated from more than 60% of the rotted drupes sampled, thus indicating that they contribute significantly to drupe rots in the regions surveyed. The most common species was B. dothidea, which was isolated from 34% of the rotted drupes. This fungus is known to be associated with a wide range of hosts in most regions of the world (Slippers et al., 2004a, 2007). Furthermore, it has long been associated with olive drupe rot, albeit under a variety of different names (Phillips et al., 2005a). In a previous study (Phillips et al. 2005a) of the same disease in Greece (based on a relatively small sample of olives), only B. dothidea was associated with the disease. According to González et al. (2006), this fungus invades the drupes through wounds caused by the olive fruit fly (Bactrocera oleae) and may even be transmitted by it. However, the present study has shown that besides B. dothidea, two species in the closely related genus Neofusicoccum (N. australe and N. vitifusiforme) are commonly associated with olive drupe rot and cause the same symptoms.

Since it was first described (Niekerk et al., 2004) N. vitifusiforme has rarely been reported. Indeed it was thought to be a weak pathogen restricted to Vitis vinifera in South Africa (Niekerk et al., 2004). However, its occurrence on 12% of the rotted olive drupes sampled in southern Italy indicates that it is more widely distributed and has a wider host range than originally thought. Neofusicoccum australe was also relatively common and was isolated from 16% of the rotted drupes. Therefore, these two species contribute significantly to drupe rot of olives.

In addition to N. vitifusiforme and N. australe, two other Neofusicoccum species were occasionally isolated. One of these clustered with known isolates of N. parvum. This species is common on many woody hosts including Vitis, Actinidia and Populus species (Pennycook & Samuels, 1985) and thus it seems to be a common, widespread, plurivorous pathogen. On some hosts (e.g. Vitis vinifera) it is known to be an aggressive pathogen (Niekerk et al., 2004). Because it was relatively uncommon on olives (0·2% of the isolates), it is probably of little importance on this host. The other species (N. mediterraneum), which clustered with four isolates from pistachio, was isolated from only 11 of the 1747 rotted drupes sampled.

The two olive cultivars tested differed significantly in their susceptibility to infection by B. dothidea and the Neofusicoccum species. This indicates that a certain level of resistance exists in some cultivars, which indicates that breeding for resistance in olive varieties may be possible. González et al. (2006) demonstrated that the state of maturity of the olive fruits affects their susceptibility and this must be taken into account in any testing of pathogenicity within a breeding programme.

Botryosphaeria and Neofusicoccum were the main causes of drupe rot in all provinces except Foggia, Cosenza, Matera and Palermo, where they were isolated from similar proportions of rotted drupes as the other fungi associated with drupe rot. In most provinces, the most prevalent species was B. dothidea, but in Lecce N. australe and N. vitifusiforme prevailed with less than 4% of the drupes affected by B. dothidea. In Matera, no B. dothidea was detected. These differences may be due to a number of factors including the predominant cultivars in each province, differences in climate and soil type and cultivation practices.

Judging from the frequency of isolation, it seems that B. dothidea is the most common cause of drupe rot of olives in southern Italy. This confirms the report by Phillips et al. (2005a) for the same disease in Greece and by González et al. (2006) in south-east Spain. However, N. vitifusiforme was the most aggressive species in the pathogenicity tests and N. australe also caused severe symptoms in artificial inoculations. Both of these two species were relatively frequent in the field and therefore they should also be taken into account in the development of disease control measures.

Acknowledgements

This work was funded by the European Regional Development Fund and Fundação para a Ciência e a Tecnologia (FCT) under project POCI/AGR/56140/2004 and by the project ‘Ricerca ed Innovazione per l’Olivicoltura Meridionale-RIOM’ under the program ‘Programma di sviluppo per il Mezzogiorno: ricerca ed innovazione tecnologica (delibere CIPE 17/2003 e B3/2003)’. A. Alves was supported by grant number SFRH/BPD/24509/2005 and A. Phillips by grant number SFRH/BCC/15810/2006 from FCT.

Ancillary