Bulletin of Entomological Research (2001) 91, 477–487 DOI: 10.1079/BER2001120

Development, survival and reproduction of black , aurantii (: ), as a function of temperature

J.J. Wang1,2 and J.H. Tsai1* 1Fort Lauderdale Research and Education Center, IFAS, University of Florida, 3205 College Avenue, Fort Lauderdale, FL 33314, USA: 2Department of Plant Protection, Southwest Agricultural University, Chongqing, 400716, People’s Republic of China

Abstract

The development, survival, and reproduction of the black citrus aphid (Boyer de Fonscolombe) were evaluated at ten constant temperatures (4, 7, 10, 15, 20, 25, 28, 30, 32 and 35°C). Development was limited at 4 and 35°C. Between 7 and 32°C, developmental periods of immature stages varied from 44.2 days at 7°C to 5.3 days at 28°C. Overall immature development required 129.9 degree-days above 3.8°C. The upper temperature thresholds of 32.3, 28.6, 29.3, 27.2, and 28.6°C were determined from a non-linear biophysical model for the development of instars 1–4 and overall immature stages, respectively. Immature survivorship varied from 82.1 to 97.7% within the temperature range of 10–30°C. However, immature survivorship was reduced to 26.3% at 7°C and 33.1% at 32°C. Mean adult longevity was the longest (44.2 days) at 15°C and the shortest (6.2 days) at 32°C. The predicted upper temperature limit for adult survivorship was at 32.3°C. Total nymph production increased from 16.3 nymphs per female at 10°C to 58.7 nymphs per female at 20°C, declining to 6.1 nymphs per female at 32°C. The estimation of lower and upper temperature limits for reproduction was at 8.2 and 32.5°C, respectively. The population reared at 28°C had the highest intrinsic rate of increase (0.394), the shortest population doubling time (1.8 days), and shortest mean generation time (9.5 days) compared with the populations reared at six other temperatures. The population reared at 20°C had the highest net reproductive rate (54.6). The theoretical lower and upper temperature limits for population development, survival and reproduction were estimated at 9.4 and 30.4°C, respectively. The biology of T. aurantii was also compared with three other citrus aphid species.

Introduction citrus encompasses 347,101 planted ha with a total of 107 million trees in the 33 citrus producing counties. The annual Citrus is one of the most important economic crops in the earning on citrus is estimated at $1.1 billion (Tsai & Wang, USA with about 500,000 ha in citrus groves mostly in 1999). The economic importance of in commercial California, Florida, Texas and Arizona. In Florida alone, citrus groves is seasonal and concomitant with new shoot growth on trees in cooler temperatures of spring and autumn. The major damage associated with the citrus *Author for correspondence aphids, including brown citrus aphid, Fax: (+1) 954 475 4125 Kirkaldy, the black citrus aphid, T. aurantii (Boyer de E-mail: jhtsai@ufl.edu Fonscolombe), the melon aphid, Aphis gossypii Glover and 478 J.J. Wang and J.H. Tsai the spirea aphid, Patch (all Hemiptera: three days throughout the study. The plants were fertilized Aphididae), however, is the transmission of citrus tristeza with a controlled release fertilizer (Osmocote, 14:14:14 virus (CTV). is found in most citrus- [N:P:K], Scotts, Marysville, Ohio) and watered as required. producing areas of the world and is the most economically Five apterous adult T. aurantii were transferred from stock important viral disease of citrus (Rocha-Pena et al., 1995). colonies to one seedling and were allow to reproduce for 4 h. Citrus tristeza virus is known to cause decline and death The adults and all but one new-born nymph were then primarily of citrus trees grafted on sour orange Citrus removed. The seedlings were caged with a plastic cage (7 aurantium L. (Rutaceae) rootstock, but some CTV isolates can 4.5 cm diameter with a nylon cloth top), and placed in cause stem pitting regardless of rootstock (Bar-Joseph et al., growth chambers (Percival, Boone, Iowa) at 4, 7, 10, 15, 20, 1989), and can result in long-term debilitation that reduces 25, 28, 30, 32 and 35°C, 70–90% r.h., and a photoperiod of yields of sweet orange and grapefruit from 5 to 45%. 14:10 h (L:D). Individual were checked twice daily Toxoptera aurantii is a polyphagous species with a for ecdysis (i.e. for life stage and developmental times) and worldwide distribution (Carver, 1978) and is reported to be a survivorship. The presence of exuviae was used to major pest of citrus in Tunisia, Italy and Peru (Talhouk, determine moulting. 1975). It is more widely distributed than T. citricida, and being more or less common with T. citricida in South Experimental procedures for adult longevity and reproduction America, Africa, India, eastern Asia and Australia but also widespread in the Mediterranean region, central America After the immatures reached adulthood and initiated and southern USA (Carver, 1978). Based on our nymphal reproduction, adult mortality and fecundity were observations, this aphid is abundant on orange jessamine recorded daily and offspring were removed from each Murraya paniculata (L.) Jack (Rutaceae) and occasionally seedling until the death of the adult. appears in citrus groves in South Florida (J.J.Wang & J.H.Tsai, unpublished data). In the past ten years, the biology Data analysis and model development and ecology of T. citricida, A. spiraecola and A. gossypii have been well documented (Komazaki, 1982, 1988; Kocourek et Developmental times, survivorship, longevity, and al., 1994; van Steenis & El-Khawass, 1995; Tsai, 1998; Tang et fecundity were subjected to analysis of variance (ANOVA) al., 1999; Tsai & Wang, 1999; Wang & Tsai, 2000; Tsai & Wang, for the effects of temperatures. General linear model 2001). However, little is known of the biology of T. aurantii, procedure (PROC GLM, SAS Institute, 1988) was used and especially about its developmental rate, temperature means were separated by Duncan‘s multiple range test thresholds, age-specific fecundity and survivorship. when significant F-values were obtained (P < 0.05). A linear Therefore, an experiment was initiated to quantify T. aurantii regression analysis (PROC REG, SAS Institute, 1988) was development, reproduction and longevity in relation to used for computing the lower developmental threshold of temperature and to provide an experimental basis for different nymphal stages (Campbell et al., 1974). developing an overall aphid population model. Development > 30 °C was outside the linear segment of the growth curve and therefore not included in the linear regression. Materials and methods Temperature-dependent developmental rates were described using the non-linear, biophysical model of Sharp Aphid source & DeMichele (1977), modified by Schoolfield et al. (1981).

Laboratory colonies of T. aurantii were established with THA  1 1  RHO25 exp  −  field-collected aphids from orange jessamine M. paniculata 298. 15  RT 298. 15  groves on the campus of Fort Lauderdale Research and r(T) = (1)  HL  11  HH  1 1 Education Center, University of Florida, Broward County, 1+−exp   +− exp    RTLT   RTHT  Florida, USA, in October 1998. Stock colonies were maintained on potted seedlings (20–40 cm tall) of orange jessamine, a preferred host for this aphid (J.J.Wang & where r(T) is developmental rate at temperature T (°K); R is J.H.Tsai, unpublished data) in an -rearing room at 25 ± the universal gas constant (1.987 cal degree1 mole–1); 1°C, 80 ± 5% r.h., and a photoperiod of 14:10 h (L:D). After a RHO25 is the developmental rate at 25°C (298.15°K) 4-month rearing period, the ensuing colonies were used for assuming no enzyme activation; HA is the enthalpy of the tests. The identity of T. aurantii was confirmed by S.E. activation of a developmental reaction that is assumed to be Halbert at the Division of Plant Industry, Florida catalysed by a rate-controlling enzyme; TL (or TH) is the Department of Agriculture and Consumer Services, Kelvin temperature at which the rate-controlling enzyme is Gainesville, Florida, USA. Voucher specimens were half active and half low- (or high-) temperature inactive; HL deposited at the collection of the Division of Plant Industry, (or HH) is the change in enthalpy associated with low- (or Florida Department of Agriculture and Consumer Services, high-) temperature inactivation of the enzyme. An SAS Gainesville, Florida, USA. program developed by Wagner et al. (1984a) using Marquardt‘s techniques was employed to fit the model to developmental data sets of different immature stages. Experimental procedures for development Cumulative frequency distributions of developmental Orange jessamine seedlings grown to 4–5 cm tall in a times for each nymphal stage were normalized using potting soil mix (50% pine bark, 40% Florida sedge peat, and median development time as the normalizing constant. A 10% sand) in pots (7.5 by 4.5 cm diameter) were used for single, temperature-independent, cumulative distribution aphid rearing, and test plants were replaced every two or was calculated as a weighted mean of the normalized Biology of Toxoptera aurantii 479

cumulative distributions, to which the cumulative Weibull where equals the number of rm to be tested, the sample size 2 distribution was fitted. of ith rm is ni. Si is the jackknife estimate of the variance for the ith r . F(x) = 1 – exp(–[(x – )/] ) (2) m The Gompertz function (Strehler, 1977) was used to where F(x) is the probability of a cohort that has completed describe the age-specific survival of female adults (lx): development at a normalized developmental time x, and , = becx , and are empirical constants. Parameters were estimated laex (6) using an SAS program developed by Wagner et al. (1984b). where a, b, and c are empirical constants; x is the age of The relationship between lifetime nymph production and adult; e equals 2.718. temperature was fitted to the modified equation of Briere et The biophysical model (equation 1) was also used to al. (1999). describe the relationship between intrinsic rate of increase y = (T – )( –T) (3) (rm) and temperature (T). Net reproductive rate (Ro) is 0 1 2 symmetric and was described by equation 3. The mean where y = lifetime nymph production, T = °C, is an generation time (MT) was fitted to a 2nd-order polynomial 0 empirical parameter, and 1 and 2 are the lower and upper function of temperature (T): lethal temperatures. The relationship between time to 50% offspring production and temperature was fitted to the MT = 0 + 1T + 2T (7) equation All models were fitted by iterative non-linear regression y = T( –T) (4) (PROC NLIN, SAS Institute, 1988) based on the Marquardt 0 2 techniques. The coefficient of determination (R2) was where y = time to 50% offspring production, T = °C, 0 is a computed as constant, and is the upper lethal temperature. Times were 2 n n 2 2 = ∑ 2 ∑ normalized by dividing by the medians, and equation 2 Ryyyy1––()()jjˆ jj– used to describe the relationship between cumulative j=1 j=1 frequency of nymph production and normalized time. The relationship between longevity and temperature was where yj is the jth observed mean value and yˆj the jth determined using equation 4. The mortality distribution of predicted value. y¯ is the mean value of all observed data. adults was determined by calculating the cumulative frequency of deaths over normalized time. The relationship Results between the cumulative frequency of deaths and normalized time was described using the model of Stinner et al. (1975). Immature development and survival F(z) = (1 – z)kz (5) Toxoptera aurantii nymphs successfully developed into where F(z) is the cumulative frequency of adult deaths, z = adults between 7 and 32°C, but failed to develop beyond the (xmax – x) / (xmax – xmin), x = the normalized time, xmin (or second instar at 4 and 35°C. For this reason, data from 4 and xmax) = the normalized time at which the first (or last) adult 35°C were excluded from the main analysis. The effects of died. temperature on the nymphal developmental periods were Life table statistics were estimated by combining highly significant (table 1, P < 0.05). The average information from the immature developmental and survival developmental period for combined nymphal stages ranged experiments with information from the adult survival and from 5.3 days at 28°C to 44.2 days at 7°C. The development reproduction experiments at different temperatures as rates of the four instars accelerated significantly with described by Hulting et al. (1990). The differences in rm increasing temperature until they reached the maximum values among populations were also analysed using developmental rates (table 1, fig. 1, P < 0.05). Longer Student-Newman-Keul sequential tests (Sokal & Rohlf, development times occurred with temperatures of > 30°C 1969) based on jackknife estimates of variance for rm (Meyer compared with the shortest developmental time at 28°C. et al., 1986). For any difference between 2 rms from the Based on a regression of development rate and all sequence, in which the rms were arrayed in order of temperatures between 7 and 28°C (the linear portion), the magnitude, to be significant at the level, it must be equal theoretical developmental thresholds were estimated to be to or greater than 3.7, 3.8, 2.9 and 4.4°C for the first, second, third and fourth instars, respectively (table 2). Toxoptera aurantii required + nnij 129.9 degree-days for a first instar to become an adult based LSR= Q s2 α[]KV. av on 3.8°C developmental threshold for overall immature 2nnij stages. The non-linear biophysical model (equation 1) provided where K is the number of rm in the set whose range is tested. The degrees of freedom equal V. The n and n were sample an excellent description of the relationship between i j developmental rate and temperature for all immature stages sizes of the 2 r s; and Q is a value from the table of the m [K.V] 2 2 as evidenced by high R values (table 2, fig. 1). The upper studentized range. Sav is the weighted average variance of rm and is calculated as follows: temperature threshold, at which high temperature began to inhibit the developmental rate of the aphid, was estimated at α 32.3, 28.6, 29.3, 27.2 and 28.6°C for the development of instars ∑()− 2 1–4 and overall immature stages, respectively. The 2 = nSii1 Sav α distribution of development at different temperatures was − ∑()ni 1 found to be a single distribution (Weibull distribution model, equation 2) as evidenced by the high R2 values (fig. 2). The 480 J.J. Wang and J.H. Tsai

Table 1. Developmental periods (days ± SE) of immature stages of Toxoptera aurantii at eight constant temperatures. Temp. °C n Instar Total of 1234immatures 7 25 9.0 ± 0.25a 8.5 ± 0.28a 9.4 ± 0.34a 17.3 ± 0.38a 44.2 ± 0.47a 10 44 3.9 ± 0.11b 5.4 ± 0.23b 4.5 ± 0.17b 6.3 ± 0.20b 20.1 ± 0.17b 15 58 2.8 ± 0.11c 2.7 ± 0.11c 2.8 ± 0.10c 3.6 ± 0.13c 11.9 ± 0.11c 20 55 2.1 ± 0.09d 2.0 ± 0.07d 1.7 ± 0.07d 2.2 ± 0.09d 7.9 ± 0.08e 25 50 1.6 ± 0.08e 1.4 ± 0.07f 1.6 ± 0.08de 1.8 ± 0.09de 6.4 ± 0.11f 28 41 1.1 ± 0.05f 1.3 ± 0.07f 1.3 ± 0.08e 1.6 ± 0.09e 5.3 ± 0.09g 30 57 1.1 ± 0.03f 1.5 ± 0.07ef 1.6 ± 0.08de 2.2 ± 0.11d 6.4 ± 0.11f 32 31 1.5 ± 0.09e 1.7 ± 0.13e 1.9 ± 0.12d 3.3 ± 0.17c 8.4 ± 0.22d F 459.8 285.2 338.5 609.7 4835.1 df 7, 353 7, 353 7, 353 7, 353 7, 353 P 0.0001 0.0001 0.0001 0.0001 0.0001

Within columns, means followed by the same letter are not significant (P > 0.05, GLM) in ANOVA (Duncan’s multiple range test).

1.0 0.8

b a 0.8 0.6

0.6

0.4 0.4

0.2 0.2

0.0 0.0

0.8

) 0.60 -1 c d

0.6 0.45

0.4 0.30 Developmental rate (day 0.2 0.15

0.0 0.00 10 15 20 25 30 35 40

Temperature (¡C) 0.20

e 0.16

0.12

0.08

0.04

0.00 10 15 20 25 30 35 40 Temperature (¡C)

Fig. 1. Developmental rate (r) of first, second, third and fourth instars (a–d), and total immature stages (e) of Toxoptera aurantii, as a function of temperature (T) in centigrade. Dashed line, linear regression for the temperature range of 7–28°C; curved line, Sharpe and DeMichele model for the range of 7–32°C. Dots are observed rates. The parameter estimates are given in table 2. Biology of Toxoptera aurantii 481

Table 2. Parameters of non-linear and linear models for temperature-dependent developmental rate of Toxoptera aurantii. Stages Poikilotherm rate function Linear modelb Instar RHO25a HAa THa HHa R2 Intercept Slope R2 c First 0.6928 13324 305.4 181185 0.98 –0.1237 0.0331 0.94 3.7 SE 0.0604 1821 0.4077 147117 0.0794 0.0042 Second 0.9960 18223 301.7 49052 0.99 –0.1212 0.0322 0.99 3.8 SE 0.1822 2348 1.8450 9443.9 0.0234 0.0012 Third 0.9138 14836 302.4 51985 0.92 –0.0901 0.0306 0.98 2.9 SE 0.3190 5131 3.5731 22024 0.0441 0.0023 Fourth 0.9033 21481 300.3 64704 0.95 –0.1199 0.0271 0.99 4.4 SE 0.2612 4157 2.1048 14441 0.0201 0.0011 First to fourth 0.2304 18387 301.7 56089 0.97 –0.0295 0.0077 0.99 3.8 SE 0.0611 3597 2.4172 16480 0.0042 0.0002 a RHO25, developmental rate at 25°C, (298.15°K) assuming no enzyme inactivation; HA, the enthalpy of activation of a developmental reaction; TH (or TL), the Kelvin temperature at which the rate-controlling enzyme is 1/2 active and 1/2 high- (or low-) temperature inactive; HH (or HL), the change in enthalpy associated with high- (or low-) temperature inactivation of the enzyme. R2, the coefficient of determination. b The linear model is for the range of 7–28°C. c The developmental threshold () = – Intercept / Slope.

1.0 1.0 a b

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0 0 1 2 3 4 0 1 2 3 4

1.0 1.0 c d

0.8 0.8

0.6 0.6

0.4 0.4 Relative frequency Relative

0.2 0.2

0.0 0.0 0 1 2 3 4 5 0 1 2 3 4

Normalized time

1.0 e

0.8

0.6

0.4

0.2

0.0 0.0 0.5 1.0 1.5 2.0 Normalized time

Fig. 2. Normalized cumulative frequency curves (solid dots) of nymphal development times of Toxoptera aurantii at various temperatures (7–32°C) and a cumulative Weibull function fitted (curves) to the weighted average cumulative frequency curve. (a) first instar, = 1.289, = 2.454, = –0.118, R2 = 0.99; (b) second instar, = 1.280, = 2.217, = –0.090, R2 = 0.99; (c) third instar, = 1.451, = 2.815, = –0.287, R2 = 0.99; (d) fourth instar, = 1.086, = 2.713, = 0.047, R2 = 0.99; (e) total immature stages, = 0.277, = 2.738, = 0.759, R2 = 0.99. 482 J.J. Wang and J.H. Tsai

Table 3. Survivorship (%) of immature stages of Toxoptera aurantii at eight constant temperatures. Temp. °C Instar Total of 1234immatures 7 72.1b 76.8b 80.3b 60.8b 26.3e 10 93.8a 97.5a 95.6a 94.2a 82.1d 15 96.7a 98.8a 98.7a 97.5a 91.9c 20 97.5a 96.6a 99.6a 99.3a 93.1bc 25 99.2a 98.6a 100.0a 100.0a 97.7a 28 98.1a 99.6a 99.3a 100.0a 96.9ab 30 94.9a 97.2a 97.0a 97.4a 87.1cd 32 74.2b 77.0b 76.8b 75.5b 33.1e F 9.05 11.1 11.4 13.9 86.6 df 7, 16 7, 16 7, 16 7, 16 7, 16 P 0.0001 0.0001 0.0001 0.0001 0.0001

There were three replicates for each temperature, and the initial number of first instar nymphs for each replicate ranged from 30 to 45. Within columns, means followed by the same letter are not significant (P > 0.05, GLM) in ANOVA (Duncan‘s multiple range test). Data were transformed to arcsine square root before statistical test; untransformed data are presented.

4 4 1.0 1.0 a b 0.8 3 0.8 3

0.6 0.6 2 2 0.4 0.4 1 1 0.2 0.2

0.0 0 0.0 0 0 10 20 30 40 50 60 0 15 30 45 60 75

8 8 1.0 1.0 c d 0.8 6 0.8 6

0.6 0.6 4 4 0.4 0.4 2 2 0.2 0.2

0.0 0 0.0 0 0 10 20 30 40 50 60 0 10 20 30 40 Reproduction (offspring per female per day) Reproduction (offspring per female 8 8 Survival (%) 1.0 1.0 e f 0.8 6 0.8 6

0.6 0.6 4 4 0.4 0.4 2 2 0.2 0.2

0.0 0 0.0 0 0 5 10 15 20 25 0 5 10 15 20 Age (days) 4 1.0 g 0.8 3

0.6 2 0.4 1 0.2

0.0 0 0 5 10 15 Age (days)

Fig. 3. Age-specific reproduction (circle and dotted line) and observed survivorship (solid dots) and simulated age-specific survivorship (curves) using the Gompertz function for Toxoptera aurantii at (a) 10°C, empirical constants of the function: a = 1.738, b = –0.448, c = 0.034, R2 = 0.97; (b) 15°C, a = 0.978, b = –0.012, c = 0.085, R2 = 0.99; (c) 20°C, a = 1.013, b = 0.004, c = 0.150, R2 = 0.98; (d) 25°C, a = 1.003, b = –0.015, c = 0.160, R2 = 0.98; (e) 28°C, a = 0.967, b = 0.015, c = 0.250, R2 = 0.95; (f) 30°C, a = 1.043, b = 0.016, c = 0.370, R2 = 0.99; (g) 32°C. a = 1.120, b = 0.041, c = 0.445, R2 = 0.98. Biology of Toxoptera aurantii 483

Table 4. Mean ± SE longevity and fecundity of Toxoptera aurantii at seven constant temperatures. 60 Temp. °C n Longevity of No. progeny a female (days) per female 50 10 30 31.7 ± 2.84c 16.3 ± 1.25d 15 56 44.2 ± 2.09a 51.6 ± 1.95b 40 20 53 36.5 ± 1.24b 58.7 ± 2.21a 25 45 21.2 ± 1.07d 52.0 ± 2.82b 30 28 30 13.7 ± 0.97e 41.2 ± 3.18c 30 32 9.9 ± 0.56ef 17.8 ± 1.13d 32 28 6.2 ± 0.41f 6.1 ± 0.53e 20 F 79.3 82.5 df 6, 267 6, 267 10

P 0.0001 0.0001 (offspring per female) Fecundity

Within columns, means followed by the same letter are not 0 significant (P > 0.05, GLM) in ANOVA (Duncan’s multiple range test).

16 b 50

a 12 40

30 8

20 4

10 Time to 50% reproduction (days) Mean longevity (days) Mean longevity 0 10 15 20 25 30 35 0 10 15 20 25 30 35 Temperature (°C)

Temperature (°C)

1.0 c 1.0 b 0.8

0.8 0.6

0.6 0.4

0.4

Cumulative frequency Cumulative 0.2

0.2 0.0 0 1 2 3 4 5 Cumulative frequency of mortality Cumulative 0.0 0.0 0.5 1.0 1.5 2.0 2.5 Normalized time

Normalized time Fig. 5. Reproductive statistics of Toxoptera aurantii as a function of temperature (a, b) or normalized time (c). (a) Total number of Fig. 4. Adult longevity of Toxoptera aurantii as a function of (a) nymphs produced per female (equation 3; 0 = 0.415, 1 = 8.182, 2 2 temperature – equation 4; 0 = 0.147, 2 = 32.257, R = 0.91, and 2 = 32.478, R = 0.98); (b) time to 50% reproduction (equation 4; (b) normalized time based on the cumulative frequency of adult = 0.055, = 32.371, R2 = 0.92); (c) cumulative frequency of 0 2 mortalities – equation 5; xmax = 2.115, xmin = 0.168, = 1.765, k = nymph production (equation 2; = 1.301, = 1.363, = –0.014, 2.132, R2 = 0.99). R2 = 0.99). 484 J.J. Wang and J.H. Tsai

Table 5. Comparison of life table parameters of Toxoptera aurantii at seven constant temperatures.

Temp. °C nrm (95% CI) Ro MT DT 10 30 0.085d (0.067, 0.103) 13.42 30.8 8.2 15 56 0.181c (0.159, 0.203) 47.69 21.5 3.9 20 53 0.266b (0.235, 0.298) 54.60 15.1 2.6 25 45 0.334ab (0.293, 0.376) 50.77 11.8 2.1 28 30 0.394a (0.333, 0.455) 40.13 9.5 1.8 30 32 0.274b (0.219, 0.329) 15.52 10.1 2.6 32 28 0.061e (0.046, 0.076) 2.02 11.5 11.4 n, Number of females in analysis. rm, Jackknife estimate of the intrinsic rate of increase (per capita rate of population growth). Within this column, the values with the same letters are not significantly different (P > 0.05). Ro, net reproductive rate. MT, mean generation time (in day). DT, doubling time (in day) for population.

parameters were significantly different from 0 (P < 0.05) 0.4 a based on non-overlap of approximate 95% confidence ) m r intervals. Overall, the adult stage began at the normalized 0.3 age of 0.8, or at 80% of the median time of development, and was completed at the normalized age of 1.3, or 130% of the median developmental time (fig. 2e). 0.2 The survivorship for immature stages of T. aurantii varied significantly across the range of temperatures tested (P < 0.1 0.05) and the stages of development (table 3). Extremely low and high temperatures (< 7°C and > 32°C) had a detrimental Intrinsic of increase ( rate effect on the survival of immature stages. The aphids reared 0.0 at 25°C exhibited the highest survivorship of 97.7%, whereas the lowest of 26.3% was recorded at 7°C. The nymphal mortality was generally low (< 10%) among instars 1–4 over 60 the range of 10–30°C (table 3). b

) Adult longevity and reproduction 0 45

R All tested aphids failed to reproduce at 7°C (n = 25). Patterns of adult survivorship and reproduction at 10–32°C 30 are shown in fig. 3. No adults died until 6–11 days after reaching adulthood at all temperatures except at 32°C. The 15 longest individual lifespans were 55, 64, 55, 28, 21, 17 and 11 days at temperatures from 10 to 32°C, respectively. The

Net reproductive rate ( rate Net reproductive Gompertz function (equation 6) gave an excellent fit to the 0 data set of age-specific survival for all test temperatures (fig. 3). Mean adult longevity was significantly different among the seven temperatures (P < 0.05), and was longest (44.2 days) at 15°C and shortest (6.2 days) at 32°C. Low temperature (10°C) caused a decline in adult longevity (table 30 c 4). This pattern was well described using equation 4 as 2

, days) evidenced by the higher R (fig. 4a). It predicted that the

MT upper temperature limit for adult survivorship was 32.3°C. 20 The distribution of adult mortalities was also well described using equation 5 (fig. 4b), with all parameters being different from 1.0 based on non-overlap of 95% confidence intervals. 10 The cumulative frequency of adult mortalities reached 1.0 at a normalized age of 2.1. Both the timing and magnitude of reproduction varied

Mean generation time ( Mean generation according to temperature (fig. 3, P < 0.05). Most females 0 10 15 20 25 30 35 began to produce within 24 h after reaching adulthood when the insects were exposed to > 15°C. Peak reproduction Temperature (°C) occurred early, then declined as females aged. The reproductive periods of most individuals lasted throughout Fig. 6. Life table statistics of Toxoptera aurantii as a function of adulthood within the temperature range of 20–28°C, temperature. (a) Intrinsic rate of increase (r ) (equation 1; m whereas the insects exposed to 10, 15, 30 and 32°C had RHO25 = 0.346, HA = 8860, TL = 282.5, HL = –67541, TH = 303.5, 2 relatively long post reproduction periods of 6.1, 5.6, 2.2 and HH = 224369, R = 0.99); (b) Net reproductive rate (R0) (equation 2 1.8 days, respectively (fig. 3). Extreme low and high 3; 0 = 0.419, 1 = 8.572, 2 = 32.212, R = 0.97); (c) Mean temperatures had adverse effects on adult fecundity (table generation time (MT) (equation 7; 0 = 58.385, 1 = –3.322, 2 = 0.058, R2 = 0.99). 4). The relationship was curvilinear, with total nymphal Biology of Toxoptera aurantii 485 reproduction increasing from 16.3 nymphs per female at 10°C to 58.7 at 20°C, then declining to 6.1 nymphs per female 0.20 at 32°C. The relationship between total nymphal reproduction and temperature was well described by a

1 0.16 equation 4 (fig. 5a). The lower and upper temperature limits Ð for reproduction were estimated as 8.2 and 32.5°C, respectively. Time to 50% nymphal reproduction increased 0.12 as temperature increased from 10°C (13.6 days) to 15°C (15.3 days), and then declined in a curvilinear fashion as temperature increased to 32°C (2.6 days). Equation 4 gave a 0.08 good fit to the data set over the temperature range of 10–32°C, and the upper temperature limit was estimated at

32.4°C (fig. 5b). The relationship between the cumulative ) (day rate Development 0.04 frequency of nymph production and normalized time was well described using equation 2 (fig. 5c). 0.00 Life table parameters

The intrinsic rate of increase (rm), net reproductive rate (Ro), mean generation time (MT), and population doubling time (DT) of T. aurantii were calculated for the populations at 60 seven temperatures (table 5). The effects of temperature on b the life table parameters were significant (P < 0.05). rm 50 increased from 0.085 at 10°C to 0.394 at 28°C, then decreased to 0.061 at 32°C (table 5). The biophysical model (equation 1) 40 fitted the data well as evidenced by the high R2 (fig. 6a), with estimated lower and upper temperature limits of 9.4 and 30 30.4°C, respectively. R0 was symmetric and was well described by equation 3 (fig. 6b), with estimated lower and upper temperature limits of 8.6 and 32.2°C, respectively. 20 Mean generation time decreased from 30.8 days at 10°C to 9.5 days at 28°C, then increased to 11.5 days at 32°C. It was 10

well described by a quadratic temperature function (offspring per female) Fecundity (equation 7) (fig. 6c). The shortest population doubling time 0 (1.8 days) was calculated at 28°C and the longest (11.4 days) at 32°C (table 5).

0.5 Discussion Temperature is an important physical environmental c 0.4

) m variable affecting the rates of development, reproduction, r and survival of aphids (Barlow, 1962; Dixon, 1987). Accurate description of the relationship between temperature and life 0.3 history parameters is thus essential to the formulation of phenological models or studies of population dynamics of insects. Our results clearly demonstrated the effects of 0.2 temperature on the nymphal development, survivorship, longevity and fecundity of T. aurantii. The relationship between temperature and development 0.1

in insects is linear over most of the normal operating, middle Intrinsic of increase ( rate range of temperature, but becomes sigmoid over the whole temperature range that permits development (Andrewartha 0.0 & Birch, 1954; Liu et al., 1995; Gilbert & Raworth, 1996). Liu 10 15 20 25 30 35 et al. (1995) have shown, using all studies available in the ° literature, that the instantaneous rate of development Temperature ( C) remained the same under constant and fluctuating temperatures. Therefore, the development-rate model derived from the constant temperature data in this study can be used to estimate the developmental time of this insect Fig. 7. Comparison of developmental rates for (a) combined immature stages (b), total number of nymphs produced per under natural conditions for the purpose of developing a female and (c) intrinsic rate of increase among three citrus aphid management strategy. species. Toxoptera aurantii (dots and solid lines), T. citricida The rate of development of T. aurantii was faster than that (hexagons and dotted lines, data from Tsai & Wang (1999)), and of A. spiraecola within the range of 10–32°C, and T. citricida Aphis spiraecola (circles and dashed lines, data from Wang & Tsai within 10–28°C, but slower at 30 and 32°C (fig. 7a) (Tsai & (2000)). 486 J.J. Wang and J.H. Tsai

Wang, 1999; Wang & Tsai, 2000). Based on day-degree theory, Our study has shown that T. aurantii could thrive well T. aurantii required 129.9 degree-days above 3.8°C to during the winter months in most citrus growing areas and complete immature development (table 2). Low temperature then significantly increase in population size in the spring thresholds of T. citricida and A. spiraecola obtained from the when the temperature increases above 15°C. Although T. constant temperature experiments (6.3 and 2.3°C, aurantii is a less efficient vector of citrus tristeza virus in respectively) (Tsai & Wang, 1999; Wang & Tsai, 2000) are comparison with other citrus aphids, the greater population somewhat lower than those from the fluctuating temperature growth rate of T. aurantii (fig. 7) will enable it to maintain experiments (7.4 and 6.7°C, respectively) (Komazaki, 1988). higher population levels to compensate for its low To date, no developmental data on T. aurantii at fluctuating transmission efficiency. In addition, Tang & Yokomi (1995) temperatures are available for comparison. reported that T. aurantii is a suitable host for three important It is well known that insects reared at temperatures above parasitoids of citrus aphids, including Aphelinus spiraecolae the upper thresholds develop slower than insects reared Evans & Schauff, A. gossypii Timberlake (both Hymenoptera: under more favourable conditions (Sharpe & DeMichele, Aphelinidae), and Lysiphlebus testaceipes (Cresson) 1977). Our data showed a similar effect. More specifically, (Hymenoptera: Braconidae). Moreover, several Aphelinus the duration of the fourth instar was relatively longer than and Lysiphlebus species are known to attack T. citricida, the those of the other three instars, especially at unfavourable, most efficient vector of citrus tristeza virus. Dixon (1998) high constant temperatures, and the instantaneous rate of stated that the global warming could result in a change in development slowed down gradually as the time the insect the distribution of species and the appearance of major pest spent at increased temperatures (table 1, fig. 2). This may be aphids in countries where they have not previously been caused by major physiological changes occurring during recorded. Therefore, the distribution range of T. citricida development from the fourth nymphal instar to the adult could be expected to increase in the future. The wide stage. Similar phenomena were also reported for three other distribution and greater population growth ability of T. citrus aphid species (van Steenis & El-Khawass, 1995; Tsai & aurantii may play an important role in sustaining parasitoid Wang, 1999; Wang & Tsai, 2000). The upper temperature populations for other citrus aphids in citrus groves. It is threshold for various stages of T. aurantii derived from the known that the development of aphid parasitoids is non-linear biophysical model ranged from 27.2 to 32.3°C dependent on temperature (Hågvar & Hofsvang, 1991), and (table 2, fig. 1). Empirically, this appears quite realistic and aphid parasitoids are expected to have rather higher the values are similar to those estimated for T. citricida (31.2° temperature thresholds than their first hosts of the season C), and A. spiraecola (c. 32°C) (Tsai & Wang, 1999; Wang & (Gilbert & Raworth, 1996). The relationships between citrus Tsai, 2000). aphids and their parasitoids therefore need to be evaluated Adult longevity and nymphal reproduction of T. aurantii further in the field. The present study has demonstrated that were correlated with temperature and age (table 4, fig. 3). T. aurantii can actually survive and reproduce over the Toxoptera citricida and A. spiraecola showed a similar pattern temperature range of 7 to 32°C, thus all the data should be of declining longevity with temperature within the 10–32°C useful in other citrus growing areas in the world. range. However, low temperature (10°C) caused a decline in female longevity of T. aurantii (table 4). A comparison of total Acknowledgements nymphal reproduction per female of T. aurantii, T. citricida and A. spiraecola (fig. 7b) indicated that nymphal Appreciation is extended to Mr J. Badmin and two reproduction was not a linear function of adult age but was anonymous reviewers for their critical and helpful review of dependent on temperature. this manuscript. This research was supported by the Florida Fertility life tables are appropriate for the study of the Agricultural Experiment Station, and approved for dynamics of populations, especially , as publication as Journal Series No. R-08213. an intermediate process for estimating parameters related to population growth potential. The r has been suggested as m References the most useful statistic for measuring the influence of various environmental factors on population growth of Andrewartha, H.G. & Birch, L.C. (1954) The distribution and aphids (Wyatt & White, 1977), or for comparing the potential abundance of . Chicago, The University of Chicago rate of increase of different species reared under the same Press. conditions (Leather & Dixon, 1984), because it incorporates Bar-Joseph, M., Marcus, R. & Lee, R.F. (1989) The continuous developmental times as well as survival and reproduction challenge of citrus tristeza virus control. Annual Review of parameters (Messenger, 1964). The rm of T. aurantii was Phytopathology 27, 291–316. higher than that of A. spiraecola over the temperature range Barlow, C.A. (1962) The influence of temperature on the growth of 10–32°C, and that of T. citricida over the range of 10–28°C of experimental populations of Myzus persicae (Sulzer) and whereas the rm of T. citricida was somewhat higher at Macrosiphum euphorbiae (Thomas ) (Aphididae). Canadian temperatures of 30 and 32°C (fig. 7c) (Tsai & Wang, 1999; Journal of Zoology 40, 145–156.

Wang & Tsai, 2000). The rm values of T. aurantii were much Briere, J.F., Pracros, P., Le Roux, A.Y. & Pierre, J.S. (1999) A lower than those of A. gossypii within the 10–30°C range novel rate model of temperature-dependent development (Kocourek et al., 1994; van Steenis & El-Khawass, 1995). The for arthropods. Environmental Entomology 28, 22–29. model predicted that the maximum rm of T. aurantii equalled Campbell, A., Frazer, B.D., Gilbert, N., Gutierrez, A.P. & 0.385 females per female per day at 27.9°C (fig. 6a), whereas Mackauer, M. (1974) Temperature requirements of some rm equalled 0.298 at 24.5°C for A. spiraecola, and 0.385 at aphids and their parasites. Journal of Applied Ecology 11, 28.8°C for T. citricida (Tsai & Wang, 1999; Wang & Tsai, 2000). 431–438.

Komazaki (1982) reported that the maximum rm of A. Carver, M. (1978) The black citrus aphids, Toxoptera citricidus gossypii was achieved at 22–23°C in Japan. (Kirkaldy) and T. aurantii (Boyer de Fonscolombe) Biology of Toxoptera aurantii 487

(Homoptera: Aphididae). Journal of the Australian rate model based on absolute reaction-rate theory. Journal of Entomological Society 17, 263–270. Theoretical Biology 88, 719–731. Dixon, A.F.G. (1987) Parthenogenetic reproduction and the rate Sharpe, P.J.H. & DeMichele, D.W. (1977) Reaction kinetics of of increase in aphids. pp. 269–287 in Minks, A.K. & poikilotherm development. Journal of Theoretical Biology 64, Harrewijn, P. (Eds) Aphids, their biology, natural enemies and 649–670. control. Vol. A. Elsevier, The Netherlands. Sokal, R.R. & Rohlf, F.J. (1969) Biometry – the principles and Dixon, A.F.G. (1998) Aphid ecology. 2nd edn, 300 pp. London, practice of statistics in biological research. 253 pp. San Chapman & Hall. Francisco, Freeman. Gilbert, N. & Raworth, D.A. (1996) Insects and temperature – a Stinner, R.E., Butler, G.D., Bacheler, J.S. & Tuttle, C. (1975) general theory. Canadian Entomologist 128, 1–13. Simulation of temperature-dependent development in Hågvar, E.B. & Hofsvang, T. (1991) Aphid parasitoids population dynamics models. Canadian Entomologist 107, (Hymenoptera: Aphidiidae): biology, host selection and use 1167–1174. in biological control. Biocontrol News and Information 12, Strehler, B.L. (1977) Time, cells and aging, VI. Ultimate effects of 13–41. cellular aging, pp. 103–125 in Mortality – a review of theories Hulting, F.L., Orr, D.B. & Obrycki, J.J. (1990) A computer of mortality. 2nd edn. New York, Academic Press. program for calculation and statistical comparison of Talhouk, A.S. (1975) Citrus pests throughout the world. Citrus, intrinsic rates of increase and associated life parameters. Ciba-Geigy Agrochemicals Technical Monograph 4, 51–54. Florida Entomologist 73, 601–612. Tang, Y.Q. & Yokomi, R.K. (1995) Temperature-dependent Kocourek, F., Havelka, J., Berankova, J. & Jarosik, V. (1994) development of three hymenopterous parasitoids of aphids Effect of temperature on development rate and intrinsic (Homoptera: Aphididae) attacking citrus. Environmental rate of increase of Aphis gossypii reared on greenhouse Entomology 24, 1736–1740. cucumbers. Entomologia Experimentalis et Applicata 71, Tang, Y.Q., Lapointe, S.L., Brown, L.G. & Hunter, W.B. (1999) 59–64. Effects of host plant and temperature on the biology of Komazaki, S. (1982) Effects of constant temperatures on Toxoptera citricida (Homoptera: Aphididae). Environmental population growth of three aphid species, Toxoptera citricida Entomology 28, 895–900. (Kirkaldy), Aphis citricola Van der Groot and Aphis gossypii Tsai, J.H. (1998) Development, survivorship, and reproduction Glover (Homoptera: Aphididae) on citrus. Applied of Toxoptera citricida (Kirkaldy) (Homoptera: Aphididae) on Entomology and Zoology 17, 75–81. eight host plants. Environmental Entomology 27, 1190–1195. Komazaki, S. (1988) Growth and reproduction in the first two Tsai, J.H. & Wang, K.H. (1999) Life table study of brown citrus and summer generations of two citrus aphids, Aphis aphid (Homoptera: Aphididae) at different temperatures. citricola van der Goot and Toxoptera citricidus (Kirkaldy) Environmental Entomology 28, 412–419. (Homoptera: Aphididae), under different thermal Tsai, J.H. & Wang, J.J. (2001) Effects of host plants on biology conditions. Applied Entomology and Zoology 23, 220–227. and life table parameters of Aphis spiraecola (Homoptera: Leather, S.R. & Dixon, A.F.G. (1984) Aphid growth and Aphididae). Environmental Entomology 30, 44–50. reproductive rates. Entomologia Experimentalis et Applicata van Steenis, P.J. & El-Khawass, K.A.M.H. (1995) Life history of 35, 137–140. Aphis gossypii on cucumber: influence of temperature, host Liu, S.S., Zhang, G.M. & Zhu, J. (1995) Influence of temperature plant and parasitism. Entomologia Experimentalis et Applicata variations on rate of development in insects: analysis of 76, 121–131. case studies from entomological literature. Annals of the Wagner, T.L., Wu, H.I., Sharpe, P.J.H., Schoolfield, R.M. & Entomological Society of America 88, 107–119. Coulson, R.N. (1984a) Modeling insect development rate: a Messenger, P.S. (1964) The influence of rhythmically fluctuating literature review and application of a biophysical model. temperatures on the development and reproduction of the Annals of the Entomological Society of America 77, 208–225. spotted alfalfa aphid, Therioaphis maculata. Journal of Wagner, T.L., Wu, H.I., Sharpe, P.J.H. & Coulson, R.N. (1984b) Economic Entomology 57, 71–76. Modeling distributions of insect development time: a Meyer, J.S., Ingersoll, C.G., McDonald, L.L. & Boyce, M.S. literature review and application of the Weibull function. (1986) Estimating uncertainty in population growth rates: Annals of the Entomological Society of America 77, 475–484. jackknife vs. bootstrap techniques. Ecology 67, 1156–1166. Wang, J.J. & Tsai, J.H. (2000) Effects of temperature on the Rocha-Pena, M.A., Lee, R.F., Lastra, R., Niblett, C.L., Ochoa- biology of Aphis spiraecola (Homoptera: Aphididae). Annals Corona, F.M., Garnsey, S.M. & Yokomi, R.K. (1995) Citrus of the Entomological Society of America 93, 874–883. tristeza virus and its aphid vector, Toxoptera citricida. Wyatt, I.J. & White, P.F. (1977) Simple estimation of intrinsic Threats to citrus production in the Caribbean and Central increase rates for aphids and tetranychid mites. Journal of and North America. Plant Disease 79, 437–445. Applied Biology 14, 757–766. SAS Institute (1988) SAS/STAT user’s guide. release 6.03 edn. 1028 pp. Cary, North Carolina, SAS Institute. Schoolfield, R.M., Sharpe, P.J.H. & Magnuson, C.E. (1981) (Accepted 21 June 2001) Nonlinear regression of biological temperature-dependent © CAB International, 2001