Fungal pathogens of mango (Mangifera indica L.) inflorescences1,2

This is the first comprehensive study to identify fungal pathogens of mango (Mangifera indica L.) inflorescences in Puerto Rico. A total of 452 mango inflorescences were collected from four cultivars at seven developmental stages during two blooming seasons. Samples were gathered from the germplasm collection at the Agricultural Experiment Station of the University of Puerto Rico in Juana Díaz, Puerto Rico. Eight different symptoms were observed: cankers, flower abortion, powdery mildew, rachis necrotic lesions, rachis soft rot, tip blight, vascular wilt, and insect perforations with necrotic borders. Necrosis was the most prevalent symptom (47%), followed by powdery mildew (19%) and tip blight (6%). Symptoms of malformation were never observed in the field. Using a modified Horsfall and Barratt scale, data on all mango cultivars pooled from two blooming seasons showed that the full bloom stage, the last inflorescence developmental stage (G), displayed the highest mean disease severity (42.67%). This severity value was significantly higher than those of the other developmental stages evaluated (P<0.05). Early inflorescence developmental stages were asymptomatic or showed the lowest percentage of disease severity. An ANOVA was performed to compare disease severity among all mango cultivars regardless of developmental stage. Results showed that there were significant differences (P<0.05) between mean disease severity of cultivars ‘Parvin’ and ‘Haden’. Mean disease severity was higher in ‘Haden’ (20%) when compared to ‘Parvin’ (10.7%). There were no statistical differences in mean disease severity between cultivars ‘Irwin’, ‘Keitt’ and ‘Parvin’, or between ‘Irwin’, ‘Haden’ and ‘Keitt’. In addition to the powdery mildew caused by Pseudoidium anacardii, 26 genera of fungi, mainly of Ascomycetes, 1Manuscript submitted to Editorial Board 18 June 2019. 2This work was supported by the USDA National Institute of Food and Agriculture, Hatch project H-425. The authors thank agronomists Víctor M. González, Luis E. Collazo and Lorena L. Simbaña for their assistance in the development of this research. 3Professor, Department of Agro-environmental Sciences, University of Puerto RicoMayagüez, Mayagüez, P.R. 00680; e-mail: lydiai.rivera@upr.edu. 4Ex-graduate student, Department of Agro-environmental Sciences, University of Puerto Rico-Mayagüez. 5Professor, Department of Agro-environmental Sciences, University of Puerto Rico, Agricultural Experiment Station, Juana Díaz, P.R. 0079. 6Research Associate, Department of Agro-environmental Sciences, University of Puerto Rico-Mayagüez. 7Research Plant Pathologist, USDA-ARS-Tropical Agriculture Research Station, 2200 P.A. Campos Avenue Ste. 201, Mayagüez, Puerto Rico 00680-5470. 140 RiveRa-vaRgas et al./Fungi oF Mango inFloRescences were identified from a total of 569 fungal isolates, from symptomatic and asymptomatic inflorescences. The most common fungi were: Diaporthe spp. (29%), followed by members of the Botryosphaeriaceae (16%), Curvularia spp. (11%) and Fusarium spp. (11%). Many fungal pathogens identified in this study were isolated from asymptomatic tissue, occurring as endophytes or latent pathogens: A. alternata, various members of the Botryosphaeriaceae, C. gloeosporioides complex, Cladosporium spp. and F. decemcellulare. Thus, the use of protectant fungicides will not be as effective as systemics in their control. Correct identification of fungal pathogens affecting mango inflorescences is important when quarantine regulations are applied. In addition, this information will facilitate the development of better management strategies in mango orchards.


INTRODUCTION
Mango (Mangifera indica L.) ranks as one of the most important traded tropical fruits (FAO, 2020). Worldwide, approximately 2 billion ton were exported during 2019, where India, China and Thailand dominated the market, producing up to 35 million ton of fruits (FAO, 2020). Even though mangoes have been grown in Puerto Rico since about 1750, it was not until 1948 that the Agricultural Experiment Station of the University of Puerto Rico (UPR-AES) started a program of mango improvement with the introduction and testing of over 150 cultivars (Morton, 1987). The most important commercial cultivar produced is 'Keitt' (80%). The other 20% includes cultivars such as 'Palmer', 'Parvin', 'Tommy Atkins' and 'Haden' (Alvarado et al., 2004). Ninety percent of mango production in Puerto Rico is exported, mostly to Europe (80%) and the USA (10%). Only 10% is available for local consumption (Morton, 1987;USDA-NAAS, 2014).
During 2018, among the Caribbean Greater Antilles, Haiti was a major producer of mangoes, mangosteens and guavas with 642,880 ton, followed by Cuba (391,111 ton) (FAO, 2020). In Puerto Rico, mango is an important fruit with a market value of $25 million (Puerto Rico Department of Agriculture, 2015) ( Figure 1). According to the most recent data from the Census of Agriculture for Puerto Rico, 1,263 ha (3,120 acres) were used for mango production in 2012 (USDA-NASS, 2014). On the island, mango fruit enterprises face major constraints. Besides the impact of devastating atmospheric disturbances such as Hurricane María, which caused $11.2 million in losses (Gómez, 2018), disease and insect pests can decimate production and consequently, exports (Alfaro, 2010). Another limitation, not related to production, is increased competition from other mango exporting countries that forces farmers to seek ways to improve fruit appearance and quality, and to reduce production costs.
Despite the relative economic importance of mangoes worldwide, and to Puerto Rico in particular, knowledge of pathogens in inflorescence tissue is practically non-existent. Very few studies have been conducted to understand fungal pathogen species and insect pest pop-ulations affecting mango inflorescences (Lonsdale and Kotzé, 1993;Prakash, 2003;Ramos et al., 1991;Ploetz, 2003;Slippers et al., 2004). More important, little is known about the mycobiota and their interactions with other pathogens and insect pests. Lonsdale and Kotzé (1993) recognized different diseases that affect mango inflorescence in South Africa. These are blossom blight and spots, malformation and powdery mildew. Curved and necrotized peduncles with tip die-back characterize blossom blight disease (Ploetz, 2003). A complex of fungal pathogens has been implicated in the disease, including a group of species belonging to the Botryosphaeriaceae (Lonsdale and Kotzé, 1993;Ploetz, 2003;Slippers et al., 2005). Another important mango inflorescence disease is blossom spot caused by Alternaria spp. In India, Alternaria tenuissima and A. alternata caused significant decreases in fruit set (Prakash, 2003). In Africa, A. alternata was reported infecting panicles besides reducing fruit set (Cronje et al., 1990cited by Ploetz, 2003. In Australia, Colletotrichum gloeosporioides and C. gloeosporioides var. minor were reported as causal agents of blossom blight (Ploetz, 2003). In Homestead, Florida, USA, besides C. gloeosporioides, C. acutatum was reported infecting mango flowers and panicles (Rivera et al., 2006). Currently, C. acutatum and C. gloeosporoides are considered species complexes (Damm et al., 2012;Weir et al., 2012). Powdery mildew, caused by Pseudoidium anacardii (syn. Oidium mangiferae and teleomorph: Erysiphe quercicola), is another major disease of mango inflorescence (Johnson, 1991;Ploetz, 2003;Prakash, 2003). In Sinaloa, Mexico, Pseudoidium anacardii has been reported affecting mango inflorescence (Félix-Gastélum et al., 2013). No teleomorph of this fungus has been described. Necrosis is observed on affected panicles with few or no fruit (Ploetz, 2003). In South Africa, the first symptoms of the disease appeared two to three weeks after 20% of the inflorescence attained the red-colored to red-open stage of development in mango cv. 'Tommy Atkins'. These authors reported 80 to 90% crop losses caused by this disease (Schoeman et al., 1995). In Puerto Rico and Florida, complete losses occur, especially when cooler temperatures and drier conditions prevail, and no management practices are used (Toro, 1988;Ploetz, 2003). Resistance to powdery mildew appears to vary among mango cultivars, but no methodological studies have been conducted on the island.

Fungal diseases associated with mango inflorescence
Mango malformation as a biotic disorder has been disputed. Its etiology remained unclear for a century; physiological conditions, in addition to a diverse group of organisms, have been implicated in the disease (Kumar et al., 1993). Symptoms associated with malformation are shortening, thickening and branching of the inflorescences, increases in flower number and size, increases in the number of male flowers, sterility or abortion of the flowers and the development of leaves within the inflorescence (Marasas et al., 2006;Ploetz, 2003;Prakash, 2003;Freeman et al., 2014). It has been demonstrated using Koch's postulates that Fusarium mangiferae (= F. subglutinans, formerly F. moniliforme var. subglutinans), isolated from vegetative shoots and floral tissue, is the causal agent of malformation in South Africa, Egypt, Israel and Florida, USA (Marasas et al., 2006). An interaction between mango bud mite, Aceria mangiferae, and F. mangiferae has been established. The researchers suggested that the mites can act as a conidial vector and assist in fungal penetration (Gamliel-Atinsky et al., 2009). In South Africa, two new Fusarium species belonging to section Liseola have been associated with mango malformation (Britz et al., 2002). Worldwide, other Fusarium species have been implicated in the disease: F. sterilihyphosum was isolated from malformed tissues in South Africa and Brazil; Fusarium sp. nov. and F. proliferatum, in Malaysia; F. oxysporum, in Egypt, Mexico and India, but their pathogenicity has not been demonstrated (Haggag et al., 2010). In Mexico, a novel species, Fusarium mexicanum, was reported as the etiological agent of mango malformation (Otero-Colina, 2010). The disease has not been reported in Puerto Rico, but the mite, A. mangiferae, has been detected in mango seedlings at nurseries (Nieves-Méndez, 2005).
In Puerto Rico, a disease described as mango wither-tip caused by C. gloeosporioides was reported by Nolla (1926). These symptoms are currently described as tip blight. In 1967, Álvarez-García reported a dieback disease of mango caused by Botryodiplodia theobroma and later in 1968, as Physalospora rhodina; both are synonymous of Lasidiplodia theobromae. He also reported this fungus as the causal agent of gummosis, die-back and fruit rot of mango (Álvarez-García and López-García, 1971). Our group has reported tip blight of mango caused by different fungal species such as Diaporthe pseudomangiferae, L. theobromae, Neofusicoccum mangiferae and N. parvum (Serrato-Díaz et al., 2013a, 2013band 2014b.
Correct identification of fungi is critical to assure effective orchard disease management and is key to enforce phytosanitary regulations. Even so, robust knowledge of key inflorescence diseases is also needed to understand the dynamics of different endophytic, pathogenic and saprophytic species present in mango orchards. Thus, the goal of this research is to provide breeders, plant pathologists, farmers and integrated pest managers with the basic knowledge needed to devise sustainable management practices, adapted to our horticultural conditions, that will reduce flower losses, and, consequently, improve mango yield.

Collection of plant material
Field surveys were conducted during two mango blooming seasons to collect symptomatic and asymptomatic inflorescences of cultivars 'Keitt', 'Haden', 'Irwin' and 'Parvin'. Inflorescences were collected at seven flowering stages of development as described by Schoeman et al. (1995) from the UPR-AES Mango Germplasm Collection located in Juana Díaz, Puerto Rico. Samples were placed in plastic bags, labeled, refrigerated and processed at the Plant Pathology Laboratory of the Department of Agroenvironmental Sciences, UPR-Mayagüez. Disease symptoms were described and disease severity (%) was estimated based on the scale developed by Lonsdale and Kontzé (1993). Inflorescences were rated from 0 to 4, based on a visual scale, where 0 is an asymptomatic inflorescence, 1 equals to 1 to 25% of the diseased area, 2 equals to 26 to 50%, 3 equals to 51 to 75%, and 4 is greater than 76% of diseased area (Lonsdale and Kotzé, 1993).

Statistical analyses
Data from two surveys was consolidated into one data set and analyzed based on the percentage of diseased tissue (symptomatic) or mycelium observed at each inflorescence stage per mango cultivar. Data conversion was adjusted as suggested by Horsfall and Barratt (H-B) (1945). For this conversion the percentage range midpoint was taken directly for each estimated interval of the visual scale of Lonsdale and Kotzé (1993) described above, in which a zero value (asymptomatic) was replaced by the value of 0.001 (Table 1). Once the data was converted, means comparisons between mango cultivars and inflorescence development stages were performed through analysis of variance and the Tukey test (α = 0.05) using Infostat Statistical Program (InfoStat/ Professional, v 2017p).

Isolation of fungi from mango inflorescences
Symptomatic and asymptomatic inflorescence tissues (1 mm 2 ) were surface sterilized with 70% ethanol, 0.7% sodium hypochlorite and rinsed with de-ionized-sterile-distilled water for one minute for each treatment. Tissue sections were transferred to potato dextrose agar acidified with 25% lactic acid (APDA). For fungal identification, pure colonies were transferred to APDA. Different culture media such as carnation leaf agar (CLA), water agar (WA), oatmeal agar (OA) or cornmeal agar (CMA) were used to induce sporulation. Isolates were incubated at room temperature (approx. 26° C) for a week.

Fungal characterization
Fungal isolates were identified using taxonomic keys (Barnett and Hunter, 1998;Boerema et al., 2004;Hanlin, 1997;Leslie and Summerell, 2006;Simmons, 2007;Úrbez-Torres et al., 2011). Semi-permanent microscopic slides were prepared from pure fungal colonies and their morphological characteristics such as mycelium, production of sexual and asexual reproductive structures including conidial size and shape, ascocarps, asci and ascospores, among others structures, were examined. Fifty conidia per isolate were measured at random (length and width) for each isolate using a compound microscope (400X, Olympus, Model 40BX, Melville, NY) 8 . Morphological characterization of powdery mildew of mango was performed using a scanning electron microscope (SEM), in addition to light microscopy (Braun et al., 2002). In brief, for SEM, inflorescence tissue sections (7 to 8 mm) were fixed using 3% glutaraldehyde in 0.1M phosphate buffer pH 7.2 (Electron Microscopy Sciences, Washington, PA). After 24 h, tissues were rinsed twice for 15 min with phosphate buffer pH 7.2. Then, samples were dehydrated by a series of ascending ethyl alcohol concentrations that ranged from 10% up to 99.9%; concentrations were increased at intervals of 10% every 15 min. Tissues were rinsed with ethyl alcohol for 15 min. A critical point drying was performed for two hours using an EMS 850 (Electron Microscopy Sciences, Washington, PA). Later, samples were mounted on an aluminum stub with a 10 mm diameter carbon cover. After that, samples were concealed with gold film for 10 min using an EMS 550X (Electron Microscopy Sciences, Washington, PA). Micrographs were taken with a Scanning Electron Microscope (JSM-5410 LV Jeol Ltd. Model, Tokio, Japan) at the center of Microscopy, Department of Biology of the UPR-Mayagüez.
Genomic DNA of fungal isolates was extracted using a commercial extraction kit (DNeasy Plant Mini Kit, Qiagen, California, USA). Analysis of the ITS1-5.8-ITS2 rDNA operon was used to complement morphological characterization (White et al., 1990). Polymerase chain reaction (PCR) was used to amplify the ITS region in a reaction containing 25 μL Amplitaq Gold® PCR Master Mix (Roche, New Jersey USA), 12 pmol of each primer, 17 μL of ultrapure water (Sigma) and 20 to 30 ng of the DNA template to reach a reaction volume of 50 μL. Polymerase chain reaction products were separated by electrophoresis (Fisher Scientific, NJ) at 100V for 45 min in a 1% (w/v) agarose prepared with 1X sodium borate buffer and 4 μL ethidium bromide (1 μg/1 μl, Sigma®, St. Louis, MO) and visualized under UV light (Quantity One® 4.5 2003, BioRad Laboratory, Inc., Japan). Amplification products were purified with QIA quick Gel Extraction Kit (QIAGEN, CA).
Purified products were sequenced in both directions using commercial facilities. Once sequenced, sequences were edited and aligned with the program Sequencher® 4.9 (Gene Codes Corporation, Minnesota, USA) and compared to GenBank database. DNA sequences were deposited in GenBank (Table 2). Diaporthe sp.

Pathogenicity tests
Pathogenicity tests were conducted on healthy mango inflorescences at orchards of cultivars 'Haden' and 'Irwin' located in the UPR-AES Mango Germplasm Collection, Juana Díaz, Puerto Rico. Under field conditions, inflorescences were superficially sterilized with a solution of sodium hypochlorite 0.07% and rinsed with deionized-sterile-distilled water for one minute. Three inflorescence rachises (wounded or unwounded) were inoculated with a conidial suspension or mycelial disks, depending on the fungal isolate evaluated. Conidial suspensions were prepared by washing the surface of the colony with 50 ml de-ionized-sterile-distilled water with three drops of Tween 20. Conidia concentration was adjusted to 10 4 conidia/ml using a hemacytometer. Mycelial disks (5 mm) were removed from the edge of a fungal colony grown on APDA for a week. A total of 24 fungal isolates were evaluated (Table 3). Untreated controls were inoculated with APDA disks or de-ionized-sterile-distilled water. Inflorescences were covered with plastic bags containing a wet cotton ball to retain humidity and reduce contamination. Data of disease severity was evaluated five and eight days after inoculation (DAI).

RESULTS
A total of 452 inflorescences were evaluated during two mango blooming seasons for four mango cultivars (i.e., 'Irwin', 'Haden', 'Keitt' and 'Parvin'). Sample sizes for the first and second survey consisted of 188 and 264 inflorescences, respectively. Of these, 50% were asymptomatic (n=230) and over 31% of the inflorescences showed low disease severity or category 1 on the Lonsdale and Kotzé scale (Table 1). Only 8% of the inflorescences showed the highest disease severity belonging to category 4 (Table 1). One-fourth of the inflorescences analyzed belongs to the earlier developmental stages (stages A to C). More than 75% of the inflorescence evaluated fell into mature developmental stages (stages D to G). The majority of the inflorescence evaluated were from mango cv. 'Irwin' (n= 168; 37%) followed by 'Keitt' (n=126; 28%), 'Haden' (n= 103; 23%) and 'Parvin' (n=55; 12%).

Symptomatology
Eight different symptoms were observed in mango orchards: cankers, flower abortion, powdery mildew, rachis necrotic lesions, rachis soft rot, tip blight, vascular wilt, and insect perforations with necrotic borders (Figure 2). Symptoms of inflorescence malformation were never observed in the field during the surveys.   Over all, necrosis was the most prevalent symptom (47%), followed by powdery mildew (19%) and tip blight (6%). Fifteen percent of the inflorescences showed a combination of symptoms, especially necrosis and powdery mildew. Less than 13% of the inflorescences showed symptoms of cankers, flower abortion, insect perforations, rachis soft rot or vascular wilts. The predominant symptom was necrosis, expressed as round, irregular or ellipsoidal lesions on inflorescence rachises and flower petals ( Figure 2C to E). Powdery mildew, caused by Pseudoidium anacardii (syn. Oidium mangiferae Berthet), was first observed on the inflorescence rachises and flowers of 'Irwin' at the green-colored stage (stage D). The disease affected all four mango cultivars evaluated. The fungus completely covered the inflorescence, causing flower and fruit abortion ( Figure 2G and H). Tip blight of mango is described as a die-back caused by an array of fungal species ( Figure 2F).

Disease severity
Direct estimates of disease severity in the field ranged from 0 to >75% of the inflorescences affected. It varies among mango cultivars and inflorescence developmental stages. After data conversion of dis- ease severity to H-B values, disease severity ranged from 0.003 to 52.30% (Table 4).
Earlier inflorescence developmental stages (A and B) did not show signs of disease severity in 'Haden' and 'Parvin' (Table 4). At these early stages, disease severity varied from 5 to 21% using a modified H-B scale in 'Irwin' and 'Keitt'. Disease severity was less than 21% at developmental stages A to E. For all mango cultivars, as inflorescences mature (stages F to G) disease severity increased, with 'Haden' showing the highest numbers at red-opened (F) and full bloom (G) stages (Table  4). Cultivar 'Parvin' in comparison, showed no significant differences between developmental stages (P<0.05).
When we considered the data of all mango cultivars pooled together, the final inflorescence developmental stage (G) showed the highest mean disease severity (42.67%) using the modified HB scale (Table 5). This was significantly different from the other developmental stages. No significant differences were observed between inflorescence stages A to E, nor between stages A, C, E and F (P>0.05) ( Table  5). Developmental stages B and D showed the lowest percentage of disease severity. In addition, an ANOVA was performed to compare disease severity between all mango cultivars regardless of inflorescence developmental stage. Results showed that there were significant differences (P<0.05) between 'Parvin' and 'Haden'. Mean disease severity was higher in 'Haden' (20%) when compared to 'Parvin' (10.7%). There were no statistical differences in mean disease severity between 'Irwin', 'Keitt' and 'Parvin', or between 'Irwin', 'Haden' and 'Keitt' (Table 6). Disease severity was rated from 0 to 4, based on a visual scale developed by Lonsdale and Kontzé (1993), and converted to a midpoint as first suggested by Horsfall and Barratt (H-B) (1945). 2 Different letters mean statistical differences using Tukey tests α = 0.05. 3 Inflorescences were collected at the seven flowering developmental stages as described by Schoeman et al. (1995).

Fungal isolation and identification
A total of 569 fungal isolates from mango inflorescences were examined during the two surveys, which included 26 genera, primarily of Ascomycetes. DNA sequence analysis using ITS region of rDNA confirmed fungal morphological characterization of 36 specimens (Table  2). Specimens identified as Bipolaris sp. or Dreshlera sp. using morphology were placed within the genera Cochliobolus spp. using the ITS region of rDNA. In addition, specimens that we placed in the Order Xylariales were grouped within the genus Hypoxylon spp.
The most common fungal genus identified was Diaporthe spp. (29%), followed by members of the Botryosphaeriaceae (16%), Fusarium spp. (11%), Curvularia spp. (11%) and Cladosporium spp. (9%) (Figure 3). Diaporthe spp. were isolated from necrotic tissues of rachises and flowers, as well as from asymptomatic tissues of all mango cultivars examined. Members of Botryosphaeriaceae, which are important plant pathogens of mango, were isolated from asymptomatic as well as Disease severity was rated from 0 to 4, based on a visual scale developed by Lonsdale and Kontzé (1993), and converted to a midpoint as first suggested by Horsfall and Barratt (H-B) (1945). Data was pooled together from all inflorescence developmental stages evaluated. 2 Different letters mean statistical differences using Tukey tests α = 0.05. symptomatic tissue showing necrotic round spots and ellipsoidal lesions of rachis, pedicels and tip blight (Figure 2). In addition, they were also isolated from rachis cankers and insect perforations with necrotic borders ( Figure 2D). Profuse black, dark to light grey mycelial growth was often associated with inflorescences harboring Botryosphaeriaceae. Mycelial threads were often confused with spider's webs in the field and observation recreated after pathogenicity tests ( Figure 4A). Among the species belonging to the Botryosphaeriaceae we identified: B. dothidea, Lasidioplidia theobromae (syn. B. rhodina), Neofusicoccum parvum (syn. B. parva), N. ribis (syn. B. ribis), and N. mangiferae. This family of fungi was isolated from all mango inflorescence stages of all cultivars evaluated including asymptomatic tissues. Fusarium spp. were isolated from all inflorescence developmental stages and mango cultivars evaluated, from symptomatic and asymptomatic tissue. Fusarium decemcellulare was isolated from inflorescences with vascular wilt, necrotic margins of flowers, rachises and pedicels; and insect perforations with necrotic borders. It was also isolated from asymptomatic tissue. Fusarium solani was isolated from long ellipsoidal lesions of flowers and small necrotic spots of rachises, as well as from asymptomatic flowers. Fusarium equiseti and Fusarium oxysporum were isolated from irregular or elliptic necrotic lesions and tip blight.
All Curvularia spp. isolates were obtained from asymptomatic tissues of flowers and rachises from all mango cultivars examined. By molecular identification, two Curvularia sp. isolates were classified as Cochliobolus lunatus (Accession No. HM060592 and HM060602). Cladosporium spp. were isolated from symptomatic tissues of all mango cultivars examined and all flowering stages except for bud swell to bud break (stage A). Cladosporium spp. were associated with necrosis of rachis and flower sepals, and necrotic ellipsoidal lesions. Cladosporium spp. were often associated with other phytopathogenic fungi such as Alternaria spp., B. rhodina, Diaporthe spp. and Fusarium solani.
Alternaria spp. were isolated from flower, rachis and sepals associated with round and ellipsoidal necrotic lesions from all cultivars examined including asymptomatic tissues. Alternaria alternata was isolated from tip blight symptoms (Accession No. GU968430). In addition, A. alternata and A. tenuissima were isolated from asymptomatic tissues of flowers, rachises and sepals, often associated with Fusarium decemcellulare and Bipolaris spp.
To our surprise Colletotrichum gloeosporioides species complex (Weir et al., 2012) occurred at a very low frequency during the surveys (0.6%), even though necrotic symptoms in mango inflorescence are often attributed to this pathogen. It was associated with flower necrosis of bud-swell to bud-break stage (stage A), the first stage of development in cultivar 'Irwin'. Fungal complexes were often detected, for example, between Alternaria spp., the Botryosphaeriaceae, Curvularia sp., Diaporthe spp. and Fusarium spp.
Certain genera of plant pathogens identified occurred at very low frequencies, ranging from 0.2 to 3%, among them Bipolaris/Dreschlera spp., Cylindrocladium sp., Pestalotiopsis spp., Phoma spp., Stemphylium spp. and Verticillium sp. (Figure 3). Fourteen percent of the fungal specimens did not produce reproductive structures on culture media nor were they identified using molecular tools. These were categorized as unknown (Figure 3).
Fifty-nine percent of the fungi were isolated from asymptomatic inflorescences. Of these, 74% are important fungal pathogens. Among them, Alternaria spp., Diaporthe spp., various members of the Botryosphaeriaceae, and Fusarium spp.

Meteorological variables
Meteorological variables measured during the first survey of mango blooming season were: precipitation which averaged 1143 mm, relative humidity that ranged from 60 to 85 percent and temperatures that fluctuated from 21 to 28 °C. During the second survey, precipitation averaged 2108 mm and temperatures fluctuated from 29 to 33 °C, both variable measurements were higher than the previous year. Data of relative humidity was not available for the second survey.
Some inflorescences exhibited severe necrosis, rachis soft rots or wilting, five or eight days after inoculation on 'Haden' or 'Irwin' (Figure 4). For example, Diaporthe spp. and D. pseudomangiferae caused extensive necrotic irregular lesions, cankers, and rachis soft rot. Fusarium decemcellulare caused vascular wilt and flower abortion along the rachises.
Various fungal species such as: Alternaria sp., A. alternata, A. infectoria, B. dothidea, C. gloeosporiodes and P. sorghina were moderately pathogenic, affecting from 16 to 30% of the inflorescences. Alternaria alternata caused ellipsoidal necrotic lesions ( Figure 4E). Colletotrichum gloeosporioides complex caused ellipsoidal necrotic lesion on rachis and flower abortion ( Figure 4D). Phoma sorghina caused cankers in rachises of 'Irwin' ( Figure 4B). Isolates identified as Leptosphaerulina spp. and Phoma exigua were not pathogenic to mango inflorescences. affected inflorescence stage was full bloom (stage G) whereas early season stages such as bud swell to bud break (stage A) and mouse ear (stage B) were either asymptomatic or showed moderate symptoms ranging from 0.003 to 20.63% on the H-B scale. Antifungal compounds such as resorcinols [5-(12-cis-heptadecenyl)-resorcinol], present in the mango peel of immature fruit, could be responsible for the resistance against fungal diseases in inflorescences at early stages of development (Cojocaru et al., 1986). Another aspect to consider is that inflorescences at the full bloom stage (stage G) have been in the field longer, exposed to fungal spores, insect and scald damage, thus rendering them susceptible to pathogens.
Mango powdery mildew, P. anacardii, was observed starting at the green-colored stage (stage D) in 'Irwin', with full bloom (stage G), the most affected stage. Our findings are similar to those reported by Schoeman et al. (1995) in an epidemiological study conducted in powdery mildew of mango in South Africa. They observed powdery mildew symptoms from two to three weeks after inflorescences reached the redcolored stage (stage E) to full bloom (stage G); this last stage showed the most severe symptoms. Thus, mango inflorescences are susceptible to P. anacardii from the protected stage (stage C) to full bloom (stage G). Climatological conditions, especially cooler temperatures, are conducive to recurrent powdery mildew outbreaks in the southern part of the island.
In addition to powdery mildew, 26 fungal genera, mainly Ascomycetes, were identified as associated with these symptoms. Diaporthe (29%) and Botryosphaeriaceae (16%) were the most common fungi of mango inflorescences. Future studies should focus on the characterization of other Diaporthe spp., the most abundant genera (145 isolates) and members of the Botryosphaeriaceae (81 isolates). Botryosphaeriaceae are considered to be stress associated pathogens; for example, B. dothidea is one of the most widespread and important endophytes or latent pathogens, occurring on trees of agriculture, forestry and natural ecosystems of importance (Marsberg et al., 2017). In our study, this species was isolated from asymptomatic tissue and caused tip blight and rachis necrosis with >35% of mycelium coverage of inflorescences in 'Haden' and 'Irwin', 8 DAI. Other Botryosphaeriaceae species identified were: L. theobromae, N. mangiferae, N. parvum and N. ribis, common mango pathogens causing tip blight and extensive rachis necrosis, worldwide. We have previously reported L. theobromae, N. mangiferae and N. parvum as important fungal pathogens of mango inflorescences in Puerto Rico (Serrato-Díaz et al., 2013a, 2013band 2014a. Neofusicoccum parvum has been reported as causing mango tip blight in Australia (Slippers et al., 2005), Brazil (de Oliviera Costa et al., 2010), Italy (Ismail et al., 2013), Perú (Javier-Alva et al., 2009), South Africa (Jacobs et al., 2002) and New Zealand (Slippers et al., 2005). More recently, Lasiodiplodia iraniensis and Neofusicoccum batangarum isolated from mango tip blight were shown to be pathogenic, causing dieback to rambutan seedlings in Puerto Rico (Serrato-Díaz et al., 2020).
Sixty-four isolates belonging to different Fusarium species were identified as associated with mango inflorescences. Among those were F. decemcellulare, F. equiseti, F. oxysporum and F. solani. In 2015, we first reported that F. decemcellulare caused wilt and vascular flower necrosis in Puerto Rico (Serrato-Díaz et al., 2015). Fusarium equiseti is a cosmopolitan soil inhabitant and a common colonizer of senescent and damaged plant tissue; thus, its role as a plant pathogen should be treated cautiously (Leslie and Summerell, 2006). Fusarium oxysporum, a widely dispersed fungus, contains non-pathogenic and many pathogenic forms usually associated with vascular wilts (Leslie and Summerell, 2006). This heterogeneous species includes many forma specialis or host specific forms. Fusarium oxysporum has been reported as the predominant species associated with root rot and wilt of plantings in mango nurseries in Pakistan (Salam-Mengal et al., 2016). Fusarium solani species complex is cosmopolitan and has been recorded as a pathogen in diverse plant species. Detailed studies on the implications of Fusarium species in mango inflorescences need to be clarified. In mango, Fusarium spp. are often implicated in malformation of inflorescences and vegetative portions of the plant (Freeman et al., 2014). This symptom was not observed in the orchards and has not been reported in Puerto Rico.
Various studies have shown the importance of endophytes or latent pathogens colonizing mango tissue as a key route for disease development during fruit maturity (Slippers et al., 2005;Morales-Rondón and Rodríguez-González, 2006). Many fungal pathogens identified in this study were isolated from asymptomatic tissues, occurring as endophytes or latent pathogens: A. alternata, various members of the Botryosphaeriaceae including L. theobromae, C. gloeosporioides, Cladosporium spp. and F. decemcellulare. Thus, the use of protectant fungicides will not be as effective as systemics in their control. The majority of these fungal species are known worldwide as necrotrophs of man-go inflorescences (Ploetz, 2003). Endophytes such as Curvularia sp., which is the third most common genus isolated in this study, have been shown to provide thermal protection when growing inside plant tissues (Redman et al., 2002). According to Jumpponen (2001), under certain scenarios, dark septate endophytes are capable of forming mutualistic associations similar to those produced by mycorrhizas in roots. The ubiquitous presence of dark septate endophytes in plant tissues, besides roots, may imply a potential mutualistic nature that will provide benefits to the tree, an aspect that needs to be explored.