Epidemiology of Declinio Disease in Citrus Groves in the State of S&oPaulo,

A. Tubelis, H. S. Prates, and A. A. Salibe

ABSTRACT. A survey of declinio affected trees was conducted in 229 commercial citrus groves at the upland area of State of S%oPaulo during the 1983-1986 period. The incidence of declinio, expressed as percentage of affected plants in the groves increased continuously with the time, follow- ing an exponential mathematical expression. The increase in the incidence of the disease with time were directly proportional to the plant population of the affected orchards. The rate of spread was inversely related to the cation exchange capacity and water retention capacity of the latosol soils. Index words. Citrus, decline rates, quantitative epidemiology, disease-progress curves.

Citrus declinio was first reported visually in the affected trees include in 1975 from a commercial citrus zinc deficiency of leaves, wilting of a grove in , State of SBo Paulo, sector or entire canopy of the tree, Brazil (7). Subsequently the disease delayed flush, thin and pale foliage, was found in isolated groves far from dieback and production of water the initial focus. Now the disease oc- sprouts. Injection of water in the curs in all citrus areas of the state. trunks (1) was used to confirm the vis- The spread of declinio is of major con- ual identification. When the water cern to the growers and the agricul- uptake of the trunk was less than 5 tural regulatory agencies. By 1987 mllmin the plant was considered to be the loss of citrus plants due to declinio affected by declinio. was estimated to be about 5 million Soil effect. The effect of the soil plants. For this reason the disease is on the behavior of the declinio inci- considered one of the most serious dence was studied only for Pera sweet problems of the citrus industry of orange. From the initial 229 citrus Brazil. groves inspected, data from 129 This paper presents the results of orange groves were used for analysis. several surveys for declinio conducted The classification of the soils on which by the agricultural regulatory agen- the orchards were grown was made cies of the State of SBo Paulo in com- according to the Soil Commission (3). mercial citrus groves of the region Statistical analyses. Mathematic which is the major producer of Brazil- equations for the data were derived ian citrus. by the method of minimum square de- viation (4). The selection of equation METHODS AND MATERIALS that best fit was made by the value of Survey procedure. Inspections the coefficient of determination. The were conducted in 229 commercial cit- following types of equations were rus groves, totaling 15.3 million used to fit the data: linear, exponen- plants, in the upland area of state of tial, bi-logarithmic and semilog- SBo Paulo. The surveys for declinio arithmic. affected trees were conducted yearly RESULTS AND DISCUSSION from 1983-1986. All groves surveyed had diseased trees in various stages Time effect. Figure 1 shows that of decline. The number of declining the number of affected plants of the and healthy trees in each grove was population increased continuously recorded by trained inspectors. with time. The representation shows Declinio diagnosis. Identification clearly that declinio incidence is not a of the plants with declinio was made linear function of the time in the visually. Disease symptoms detected edafo-climatic conditions of the State Eleventh IOCV Conference

conditions of the State of SBo Paulo, the period of time twas 21.2 years. The differences in the values of the lapse of time can be due to the differ- ences in tree spacing as pointed out by Yokomi et al. (8). Population effect. The relation- ship between the incidence of the dis- ease and the tree population of groves in four agricultural regions in the State of SBo Paulo is shown in the Fig. 2. The percentage of affected trees increased with the size of the population. Thus the incidence is low in the agricultural region of , where the tree population is small and is high in the agricultural regions of RibeirBo Preto and , where the tree population is large. Lima et al. (5) found that blight incidence in some orchards of the State of SBo d Paulo, Brazil, was directly propor- I 1 1 1 4 tional to the number of blighted trees 1983 1- I 1906 1987 YEARS observed when the grove was sur- veyed 3 years earlier. Fig. 1. Relation of declinio incidence with Figure 2 also shows that for the time in 229 citrus groves in the State of Sfio same agricultural region the increase Paulo, Brazil. in declinio incidence is not a constant with time. Declinio incidence in- of SBo Paulo. Yomoki et al. (8) re- creases continuously year after year ported that the blight incidence in as seen in Fig. 1. Thus, in the same Florida was a linear function of time, agricultural region, the loss of trees the slope of the regression line was caused by the declinio will increase 3.5% per year for tree spacing of 7.6 with time and with the increase of the x 7.6 m in a Winter Garden grove. tree population. The mean percentage of affected Geographic distribution. The ge- plants in the control population in the ographic distribution and numbers of State of SBo Paulo could be expressed diseased trees in 1986 in the 229 citrus by the equation: groves in the State of Sgo Paulo is y = 1.1435 e(0.3255x) equation 1 shown in the Fig. 3 and Table 1. In the Campinas region there were where: y = percentage of affected 289,370 affected trees corresponding trees and x = number of years after to 4.3% of the total population. The 1982. The coefficient of determination diseased trees were mostly found in of this regression was 0.9985. the municipalities of Casa Branca and With the equations used by Mogi Gua~u.The most affected grove, Yokomi et al. (S), it is possible to esti- located in the municipality of Mogi mate the period of time in which the Guagii, had an incidence of 20%. disease incidence increases from 0.1% In the agricultural region of to 100%. This lapse of time was of the RibeirBo Preto, there were 299,923 order of 28.6 yr for the Winter Gar- affected trees, corresponding to 5.7% den grove, and 21.7 yr for the Avon of the total population. The diseased Park grove. Using similar procedure trees were mostly in the munici- with the equation 1 and mentioned palities of and . lapse of time for the edafo-climatic The most affected grove, with an inci- Blight and Related Declines

Fig. 2. Increase of the number of declinio affected plants in relation to the total tree popula- tion in four agricultural regions of the State of Sio Paulo, Brazil, from 1983 to 1986.

dence of 33.7%, was located in the Dark-Red Latosol sandy phase - LEa municipality of . soil or Ortho Red-Yellow Latosol - In the agricultural region of Sfio LV, which presents medium texture Jose do Rio Preto, there were 41,475 (sandy phase). This points out the affected trees, corresponding to 1.5% paramount influence of the soil tex- of the total population. The diseased ture and, consequently, the cation ex- trees were mostly in the munici- change capacity (CEC) and water palities of Mendonca and . The holding capacity (WHC) of the soils in most affected grove, with an inci- the incidence of declinio. Soil with dence of 7.2%, was located in the clay texture have higher CEC and municipality of Severinea. WHC than soils with medium tex- Soil effect. The incidence of de- ture. clinio increased continuously with Groves on Ortho Dark-Red Latosol time in all agricultural regions and in - LE soil, which are clay soils, have a all types of Latosol soils. The be- lower incidence of declinio than havior of the disease incidence with groves on Dark-Red Latosol sandy the types of Latosol soils in the State phase - LEa soil, which have medium of Sfio Paulo is shown in the Fig. 4. texture. Thus the incidence of declinio Groves grown on Ortho Dark Latosol was lower in groves on soils with - LE soil and on Latosol "Roxo" - LR higher WHC. Groves grown on Yel- soil, which is a clay soil, shows lower low-Red Latosol sandy phase - LVa incidence of declinio than groves on soil, in the agricultural region of Cam- 280 Eleventh ZOCV Conference

Fig. 3. Incidence of declinio in 229 selected citrus orchards in the State of Sio Paulo, Brazil in 1986. The numbers refer to the municipalities listed in Table 1.

pinas, are the only exception to that et al. (8) applied to observations in observation. State of Siio Paulo, it would take 21.2 The following scale of values for yr for the declinio incidence to pass incidence of declinio in citrus groves grown on Latossol soils, in decreasing order, was observed in the State of Siio Paulo: 1. Dark-Red Latosol sandy phase - LEa; 2. Ortho-Red Yellow Lastosol - LV; 3. "Roxo" Latosol - LR; 4. Ortho Dark-Red Latosol - LE; 5. Red-Yellow Latosol sandy phase - LVa. These results agree with the re- ports of Rhoads (6) and Cohen (2). Rhoads (6) observed that blight was severe on soils with low WHC and on soils with poor drainage. Cohen (2) observed that areas with high organic soil remained relatively free of blight. I The data obtained from these IL 8L tm IL I& ' exhaustive series of surveys to deter- YEARS mine the amount and rate of increase Fig. 4. Relation of the declinio incidence of declinio affected trees in citrus or- in 129 groves of Pera sweet orange gown in chards of State of Siio Paulo, Brazil, Latosol soils in the State of Sio Paulo, allows these conclusions. The declinio Brazil; LE = Ortho Dark Latosol soil; Lea = Dark-Red Latosol sandy phase soil; LR = incidence was not a linear function of Latosol "Roxo"; LV = Ortho Red-Yellow the time but instead an exponential Latosol soil; Lva = Yellow-Red Latosol sand function. Using the method of Yokomi phase soil. Blight and Related Declines 281

TABLE 1 NUMBER OF TREES AFFECTED BY DECLINIO IN 229 COMMERCIAL CITRUS GROVES IN 1986 IN STATE OF SAO PAULO, BRAZIL

Country1 No. of Country1 No. of Municipality diseased trees Municipality diseased trees

Ribeirfio Preto Campinas 13arretos 26. 2. Colina 27. Concha1 3. 28. 4. Viradouro 29. Mogi-Mirim 5. Monte Azul 30. 6. Bebedouro 31. Slo Pedro 7. Pitangueiras 32. Sta. Barbara do Oeste 8. Taidva 33. Sto. AntBnio da Posse 9. Taiacd 34. Jaguaridna 10. 35. Mogi Guacd 11. 12. Monte Alto Sorocaba 13. 36. 14. 37. Sorocaba 15. Matlo 16. Slo Jose do Rio Preto 17. 38. 18. Colombia 39. Estrela do Oeste 40. Guaraci Campinas 41. 19. Tambad 42. Palestina 20. Casa Branca 43. Sebastianbpilis 21. Sta. Cruz Palmeiras 44. 22. 45. Mendonca 23. Aguai 46. Ibira 24. Leme 47. Itajobi 25. 48. Santa Adelia

'Numbers refer to locations on map in Figure 3. from 0.1% to 100%. The rate of inci- ACKNOWLEDGEMENTS dence of the declinio incidence was di- rectly proportional to the tree popula- The authors thank CNPq - Con- tion. The most affected grove in 1986 selho de Desenvolvimento Cientifico was in the municipality of Viradouro e Tecnol6gico and FUNDUNESP - with an incidence of 33.7%. The most Fundagiio para o Desenvolvimento da affected agricultural region was UNESP, for financial support for the Ribeirgo Preto, the largest citrus research project and presentation of area of the country, with an incidence this paper, also, thanks to the many of 5.7%. The texture of the Latosol agronomists for CAT1 - Coorden- soils affected the rate of increase of adoria de Assistsncia Tecnica Inte- declinio. Latosol soils with higher gral who helped in the field survey. CEC and WHC had lower rates of de- clinio increase.

LITERATURE CITED

1. Cohen, M. 1974. Diagnosis of young tree decline, blight, and sand hill decline by measurement of water

uptake using-- gravity - injection.~ Plant. Dis. Rep. 58: 801-805. 2. ohe en, M. 1980. Nonrandom distribution of trees with citrus blight,-. p. 260-263. In Proc. 8th Conf. IOCV. IOCV, Riverside. 282 Eleventh IOCV Conference

3. Comisslo de Solos 1980. Levantamento de reconhecimento dos solos do Estado de Slo Paulo. Servico Nacional de Pesquisas AgronBmicas - CNEPA, . Boletim no. 12. 4. Ezekiel, M. and K. A. Fox 1959. Methods of correlation and regression analysis - linear and curvilinear. John Wiley & Sons, Inc. 5. Lima, J. E. O., and A. S. Borducchi 1982. Observation on citrus blight in SLo Paulo, Brazil. Proc. Fla. State Hort. Soc. 95: 72-75. 6. Rhoads, A. S. 1936. Blight - a nonparasitic disease of citrus trees. Fla. Agr. Expt. Sta. Bull. 296. 64 p. 7. Rossetti, V. V., and M. J. Beretta 1987. Declinio of citrus trees in Brazil, p. 263-272. In: Proc. Int. Symp. Citrus Canker, DeclinioIBlight and Similar Diseases. Fundacao Cargill, SLo Paulo, Brazil. 8. Yokomi, R. K., S. M. Garnsey, R. H. Young, and G. R. Grimm. 1984. Spatial and temporal analysis of citrus blight incidence in 'Valencia' orange groves in Central Florida, p. 260-269. In: Proc. 9th Conf. IOCV. IOCV, Riverside.