Appl. Entomol. Zool. 37 (4): 559–569 (2002)

Prediction of the life cycle of the west Japan type yellow-spotted longicorn , (Coleoptera: Cerambycidae) by numerical simulations

Yuya Watari, Takehiko Yamanaka, Wataru Asano and Yukio Ishikawa* Laboratory of Applied Entomology, Department of Agricultural and Environmental Biology, The University of Tokyo, Tokyo 113– 8657, Japan (Received 3 December 2001; Accepted 3 July 2002)

Abstract A temporally structured model that enables simulation of the development of the west Japan type yellow-spotted longicorn beetle, Psacothea hilaris (Pascoe), at different locations was developed. Life history parameter values in- corporated into the model were estimated by laboratory rearing experiments. To validate the present model, the devel- opment of eggs laid monthly from June 1 through November 1 was simulated under dynamic temperature and pho- toperiod conditions at Ayabe City. The individuals laid on June 1 did not enter diapause but emerged in early August of the same year. On the other hand, about 2/3 of the individuals laid on July 1, and all those laid on August 1 and September 1 entered diapause (or quiescence), and started to emerge in late May of the following year. Individuals laid on October 1 and November 1 overwintered as young larvae (1st–3rd stadia) and eggs, respectively, and the ma- jority of these emerged in late July–early August. Interestingly, the remaining individuals entered diapause in the 2nd year and emerged in June of the 3rd year. Analyses of these simulation results suggested that concentrated emergence of P. hilaris can occur twice in one year (in late May–early June and in late July–early August) at Ayabe, and this is fairly concordant with known adult prevalence at this location considering the long life-span of adults. It was also sug- gested that although P. hilaris at Ayabe has basically a univoltine life cycle with a facultative diapause, varying pro- portions of the population appear to have a bivoltine life cycle or a semivoltine life cycle depending on the meteoro- logical conditions of that year. The life cycle of P. hilaris is suggested to be very flexible and adaptive to environmen- tal fluctuations.

Key words: Psacothea hilaris, seasonal prevalence, temporally structured model, bivoltine life cycle, semivoltine life cycle

continuous stripes (Iba, 1980). INTRODUCTION These two types are also characterized by differ- The yellow-spotted longicorn beetle, Psacothea ent seasonal life cycles and photoperiodic responses hilaris (Pascoe), is widely distributed in eastern- (Iba et al., 1976; Sakakibara and Kawakami, 1992). most Asia. This species is known to have large ge- The west Japan type has its main adult prevalence ographic variation in morphology, and is classified peak in early summer, and a small second peak is into 13 subspecies (Kusama and Takakuwa, 1984). observed in autumn (Iba et al., 1976; Suzuki and One subspecies, P. hilaris hilaris, inhabits Honshu, Yoshii, 1980; Ishiishikawa, 1986). They are consid- Shikoku and Kyushu, three of the four main islands ered to enter diapause as mature larvae in response of Japan. Within this subspecies, two morphologi- to low temperatures and short daylength in autumn cal types, the “west Japan type” and “east Japan (Shintani et al., 1996b; Shintani and Ishikawa, type” have been recognized (Iba, 1980; Makihara, 1997a). On the other hand, the east Japan type does 1986). The west Japan type inhabits Kyushu, not form a sharp emergence peak, but an increase Shikoku and western parts of Honshu, and is char- in the adult population is usually found in autumn acterized by severed yellow stripes on the prono- (Iba, 1976; Iba et al., 1976; Sakakibara and Kawa- tum. In contrast, the east Japan type is found in the kami, 1992). They had been regarded to lack dia- eastern parts of Honshu, and is characterized by pause and overwinter at the egg stage (Makihara,

* To whom correspondence should be addressed at: E-mail: [email protected]

559 560 Y. Watari et al.

1986; Sakakibara and Kawakami, 1992), however, tion Shintani and Ishikawa (1997a) have shown that 1/Dϭ(TϪT )/K (1) they also enter diapause but at temperatures lower 0 than for the west Japan type. where D is the duration of a developmental stage in P. hilaris is a serious pest of mulberry and fig days, T is the constant rearing temperature, K is the trees. Control of P. hilaris is difficult because lar- thermal constant for the stage, and T0 is the lower vae bore tunnels in the trunks and are thus invisible thermal threshold for development (developmental from the outside. In addition, extensive insecticide zero). treatment of mulberry trees is not practical since Except for the egg stage, K and T0 were esti- mulberry leaves are fed to the silkworm, Bombyx mated from the information obtained from the mori. P. hilaris damages host trees heavily when above rearing experiments using a linear regression they are larvae, particularly after the 3rd stadium of 1/D against T (see Results). K and T0 for the egg (Iba, 1993). Thus, prediction of the detailed age- stage were obtained from Sakakibara (1995), who structure is indispensable to determine the best determined these values for P. hilaris from various time to control P. hilaris. locations in Japan. The values for the Kochi City The objective of the present study was to con- population (west Japan type) were used in the pres- struct a temporally structured model for prediction ent simulations. of the time of adult emergence, and to compare the Critical daylength. Shintani and Ishikawa (1998) simulation results with previous findings on adult reported that the critical daylength (CDL) for dia- prevalence in the field. Since the larval photoperi- pause induction of the west Japan type P. hilaris odic responses of P. hilaris have mainly been falls between 13.5 and 14 h at 25°C. However, they investigated using the west Japan type population did not determine the CDL at other temperatures. (Shintani et al., 1996a; Shintani and Ishikawa, In the present simulation, CDL was set as 13.75 h 1997a, b, c, 1998), in the present study, we con- (the midpoint of 13.5 h and 14 h) regardless of the structed a model for the west Japan type P. hilaris temperature, and used as a parameter that switches incorporating its daily age-structure, larval and long-day type responses and short-day type re- pupal development, and local meteorological data. sponses. Namely, when the daylength of the day t

(DLt) was greater than CDL, temperature depend- ent responses similar to those at 15L : 9D were MATERIALS AND METHODS Յ assumed to occur. Conversely, when DLt CDL, Parameter estimation responses similar to those at 12L : 12D were as- Rearing experiment. Adults of the west Japan sumed. type P. hilaris were collected from mulberry fields Meteorological parameter values incorporated in Ino Town (33.5°N, 133.4°E), Kochi Prefecture, into the simulation. The daylengths between Janu- and maintained in the laboratory at 25°C under a ary 1 and December 31 at Ayabe were calculated 15 h light–9 h dark photoregime (15L : 9D). Eggs by the sunrise and sunset times in 2001 down- were obtained as described by Shintani et al. loaded from the Hydrographic Department, Japan (1996a). Newly-hatched larvae were individually Coast Guard (http://www.jhd.go.jp/cue/KOHO/ placed in plastic Petri dishes (6 cm diam.ϫ1 cm) automail/sun_form4.htm). The daily maximum, with a piece of artificial diet (Silkmate 2S®, Nihon mean and minimum air temperatures averaged for Nosan K.K., Yokohama, Japan). Sixty to 100 lar- eight years between 1960 and 1967 at Kyoto mete- vae each were subjected to ten combinations of two orological station, which is the station nearest to photoregimes (12L : 12D and 15L : 9D)ϫfive tem- Ayabe City, were obtained from the Monthly Re- peratures (20.0, 22.5, 25.0, 27.5 and 30.0Ϯ0.1°C). port of the Japan Meteorological Agency. Molting, pupation and adult eclosion were checked Model construction daily for 180 d from the start of the experiment. Framework. The daily-structured model was de- Thermal constants and developmental thresholds veloped using Microsoft Developer Studio (version for each stage. It is well-known that the growth rate 4.0, Microsoft®) with Cϩϩ programming lan- of is proportionate to temperature within an guage. The model has 365 time steps for a one- optimum range, and can be expressed by the equa- year season, and each time step (t) corresponds to Life Cycle of West Japan Type P. hilaris 561

Fig. 1. The schematic flowchart of procedures in the simulation model. Arrows indicate the flows of the processes in one time step (one day). each day between January 1 and December 31. The riod and the mean temperature as described below. calculations were started on the day of oviposition Individuals were assumed not to suffer from any

(D0) and continued until all individuals emerged as mortality; all survived until the end of the simula- adults. During the simulation period, 100 individu- tion. These procedures are shown in Fig. 1 as a Յ Ͻ als (YSLi :0 i 100) grew uniformly at a daily rate schematic flow chart. determined by the maximum and minimum tem- Growing process dependent on the temperature. peratures of the day (see the next section) until the The stage and mode of the i-th individual were des- ϭ end of the 3rd stadium. Thereafter, the individuals’ ignated as Si ( egg, 1st, 4th, ...9th or pupa) and ϭ destiny (molting, pupation or diapause) bifurcated Mi ( normal, pupation or diapause), respectively. at the end of each stadium based on the photope- The cumulative heat unit (degree-days) was used as 562 Y. Watari et al. the index of development of an individual. The cu- ities for diapause and pupation from 100%. When mulative heat unit of i-th individual on day t (Hi,t) the average temperature of the day exceeded 30°C was obtained by or fell below 20°C, frequencies at 30°C or 20°C ϭ ϩ were adopted, respectively. After determination of Hi,t Hi,tϪ1 DHi,t (2) Ј the new mode (Mi ) as described above, the degree- Here, DHi,t, the heat unit accumulated in a day t, days (Hi,t) based on the previous lower thermal was estimated by the single triangle method (Lind- threshold T was converted to that (HЈ ) based 0Si,Mi i,t sey and Newman, 1956) with an intermediate cut- on the new threshold T Ј by the equation 0Si,Mi off (Roltsch et al., 1999; the California’s IPM Ј ϭ ϩ Ϫ Hi,t Hi,t (T0S ,M T0S ,MЈ)·Dhalf (4) Worldwide Web Site, http://www.ipm.ucdavis.edu) i i i i from the maximum and minimum temperatures of where Dhalf is the time (days) required for the indi- the day t (Tmax , Tmin ), and the lower and upper vidual i to accumulate K /2 in the developmental t t Si,Mi thermal thresholds for development (T0, Tup). Tup is stage. the temperature at which the rate of development The larvae that entered diapause did not accu- begins to decrease. The intermediate cut-off option mulate heat units unless their diapause was termi- assumes that the portion of heat unit accumulated nated. In the present model, diapause was termi- ϭ over Tup is not only ineffective but inhibitory to the nated automatically on March 1 (t 424 and 789), growth of the , and subtracts 2ϫthe excessive and the overwintered individuals pupated at the heat from the raw degree-days estimate. next molt without re-entering diapause. On the

When Hi,t of an individual in the Si stage and Mi other hand, all individuals that encountered winter mode exceeded the thermal constant specific for without entering diapause were assumed to survive the stage and mode (KS ,M ), the stage was pro- in a quiescence state. The overwintered individuals Ј i i ceeded to the next (Si ), and the surplus heat unit were also assumed not to enter diapause in early based on the previous lower thermal threshold spring, when the air temperatures start to exceed T0 (T ) was converted to that based on the new but daylength is still short, due to apparent loss of 0Si,Mi threshold (T0SЈ,M ). Overall, the cumulative heat unit sensitivity to the photoperiod (cf. Shintani and i i Ј on the first day in the next stage (Hi,t) was ex- Ishikawa, 1997c). pressed as Executing the model. Firstly, we executed simu- Ј ϭ Ϫ lations under constant temperatures and daylengths Hi,t (Hi,t KS ,M ) i i corresponding to the laboratory experiments. The ·(1ϩ(T ϪT Ј )/DH ) (3) 0Si,Mi 0Si ,Mi i,t model was recursively validated by comparison of Determination of molting, pupation or diapause. the simulation results with the results of laboratory In the present model, larval destiny to molt into rearing experiments. Secondly, simulations were the next stadium, to pupate or to enter diapause executed under the dynamic temperature and pho- was determined at the midpoint of the stadium toperiod conditions at Ayabe City, where informa- (H ϭK /2), since the sensitivities to light and tion on the prevalence of adults in 1962–1967 is i,t Si,Mi temperature are known to be highest in the former available (Iba et al., 1976). Simulations were half of each stadium (Shintani, unpublished data). started on June 1, July 1, August 1, September 1, The underlying mechanisms that induce diapause October 1 and November 1 as the date of oviposi- or pupation are not fully understood and thus can- tion. not be incorporated into the model as formulae, but the destiny of individuals was determined stochas- RESULTS tically as follows. First, the daylength of day t (DLt) was compared with CDL to select long-day type or Development of P. hilaris larvae under labora- short-day type responses. Then, the probabilities tory conditions for diapause and pupation were estimated by linear The development of the west Japan type P. hi- extrapolation of the respective frequencies that laris under different temperatures and photoperiods were experimentally obtained between 20°C and in the laboratory is shown in Fig. 2. Detailed analy- 30°C at 2.5°C intervals. The probability for larval ses of the physiological aspects of the results will molting was calculated by subtracting the probabil- be conducted and reported elsewhere. At 25°C, P. Life Cycle of West Japan Type P. hilaris 563

Fig. 2. The results of rearing experiments and those of simulations under the conditions corresponding to laboratory experi- ments. Parameter values used in this model were obtained from rearing experiments in the laboratory. Each horizontal bar in a fig- ure represents the development of an individual. hilaris larvae went through the 4th or 5th stadium The duration of each developmental stage under and pupated under a long daylength, however, laboratory conditions is shown in Table 1. The du- under a short daylength, they repeated a few more ration of a stage was shortened as the temperature molts and eventually entered diapause. Since P. hi- was increased from 20°C to 27.5°C. At 30°C, how- laris larvae grow uniformly and do not enter dia- ever, developmental rates were often smaller than pause or pupate until the end of the 3rd stadium those at 27.5°C (Table 1). Since 30°C was appar- (Shintani et al., 1996a), the 1st to 3rd stadia were ently beyond the optimum range of temperature for treated as one stage. development, Tup was set as 30°C in the present 564 Y. Watari et al.

Table 1. The durations of each stage (days) under five different temperaturesa,b

Stage 20°C 22.5°C 25°C 27.5°C 30°C

1st–3rd 24.2Ϯ2.6 (200) 20.9Ϯ1.9 (196) 17.0Ϯ1.8 (180) 14.1Ϯ1.7 (155) 11.6Ϯ1.5 (158) 4th 19.5Ϯ4.0 (200) 14.6Ϯ3.7 (183) 13.0Ϯ3.4 (160) 10.7Ϯ2.9 (123) 12.1Ϯ5.0 (105) 4th (pupation) — 29.2Ϯ3.2 (13) 24.9Ϯ3.1 (20) 21.8Ϯ2.4 (29) 21.6Ϯ4.7 (53) 5th 27.4Ϯ4.3 (182) 23.7Ϯ5.3 (144) 19.6Ϯ4.2 (85) 17.5Ϯ3.8 (68) 18.1Ϯ5.2 (35) 5th (pupation) — 31.4Ϯ6.0 (21) 23.0Ϯ4.1 (68) 20.5Ϯ3.8 (53) 21.1Ϯ5.2 (63) 6th 29.3Ϯ4.7 (80) 26.6Ϯ5.3 (52) 22.5Ϯ4.7 (22) 18.8Ϯ3.1 (41) 19.9Ϯ3.6 (15) 6th (pupation) — 30.3Ϯ9.3 (3) — — — 7th 26.3Ϯ4.4 (6) — 22 (1) 22.0Ϯ6.2 (11) 17.6Ϯ4.5 (8) 8th — — — 18 (1) — Pupa — 14.7Ϯ0.93 (40) 12.1Ϯ0.93 (88) 9.8Ϯ0.82 (85) 9.1Ϯ0.96 (116)

a Obtained from laboratory experiments disregarding the daylength. These data except for at 30°C were used to calculate the lower thermal thresholds and the thermal constants. b MeanϮSD (sample size). —: None existed.

Table 2. Parameters of the lower thermal threshold (T0) and thermal constant (K) for the development used in the simulationsa

Mode (Mi)

Stage Normal Pupation/diapause

T0 K Regression line T0 K Regression line

Egg 10.1b 95b ———— 1st–3rd 10.0 250 yϭ0.0040xϪ0.040 — — — 4th 10.3 185 yϭ0.0054xϪ0.056 7.9 435 yϭ0.0023xϪ0.018 5th 7.3 357 yϭ0.0028xϪ0.020 12.7 294 yϭ0.0034xϪ0.043 6th 7.1 385 yϭ0.0026xϪ0.019 10.9d 351d yϭ0.0029xϪ0.031d 7th 7.1c 385c yϭ0.0026xϪ0.019c 10.9d 351d yϭ0.0029xϪ0.031d 8th 7.1c 385c yϭ0.0026xϪ0.019c 10.9d 351d yϭ0.0029xϪ0.031d 9th 7.1c 385c yϭ0.0026xϪ0.019c 10.9d 351d yϭ0.0029xϪ0.031d Pupa 12.6 147 yϭ0.0068xϪ0.085 — — —

a ϭ The stadium destined to pupate or enter diapause (Mi pupation/diapause) has different K and T0. b Data obtained from Sakakibara (1995). c Parameters could not be accurately estimated from the present laboratory experiments due to the small sample size, but the sample data fit well to the regression line for the 6th stadium under the normal mode. d Parameters could not be accurately estimated due to the small sample size, but the sample data fit well to the medial line be- tween the regression lines for the 4th and 5th stadia under the pupation/diapause mode. study regardless of the stage. The data at 30°C In these cases, a regression line for the previous were excluded when calculating K and T0. stadium or a medial line between the regression Each stage of P. hilaris had different K and T0 lines for the two previous stadia were adopted for (Table 2). In addition, the final stadium that was rough estimation (see footnotes to Table 2). destined to pupate or enter diapause was shown to have different K and T0 from the same counterpart Probability for molting, pupation and diapause (Table 2). Some parameter values of K and T0 Table 3 shows the frequencies of diapause and could not be experimentally estimated due to the pupation at each stadium under various conditions small sample size of specific developmental stages. in the laboratory. The responses of the 1st to 3rd Life Cycle of West Japan Type P. hilaris 565

Table 3. The occurrence of diapause and pupation (%) at the end of each stadium under five different temperaturesa

Stadium Daylength 20°C 22.5°C 25°C 27.5°C 30°C

4th Long Diapause 00000 Pupation 0 13 22.7 37.6 40 Short Diapause 00000 Pupation 000019.1

5th Long Diapause 6.3 0000 Pupation 0 27.6 100 100 100 Short Diapause 12 15.6 7.6 2.9 15.6 Pupation 00006.7

6th Long Diapause 46.2 49.2 — — — Pupation 0 4.8 — — — Short Diapause 68.2 71.6 74.1 39.7 57.1 Pupation 00000

7th Long Diapause 87.5 100——— Pupation 0 0——— Short Diapause 100 100 95.5 73.2 46.7 Pupation 00000

8th Long Diapause 100———— Pupation 0———— Short Diapause — — 100 90.9 100 Pupation — —000

9th Long Diapause ————— Pupation ————— Short Diapause — — — 100 — Pupation — — — 0 —

a Obtained from laboratory experiments. Those larvae that neither entered diapause nor pupated molted to the next stadium. —: No larvae reached this stadium. stadia are not listed in Table 3, because P. hilaris tions are exceedingly favorable (long day and did not diapause or pupate in these stadia, as ex- Ͼ22.5°C). pected. In P. hilaris, temperature had a profound effect on the photoperiodic responses of the larvae; Simulations under constant temperatures and the effect of the photoperiod was distinct at rela- daylengths tively high temperatures (Ն25°C), but apparently The simulation results under 10 constant condi- negligible at the relatively low temperature of 20°C tions matched qualitatively with the results of cor- (Fig. 2). Under long day conditions, no larvae en- responding rearing experiments, although no indi- tered diapause at higher temperatures (25, 27.5 and vidual variations in the developmental period were 30°C), however, all larvae reared at 20°C and about exhibited in the simulations (Fig. 2). This is be- 65% of larvae reared at 22.5°C entered diapause in cause individual variations in the development the 5th, 6th or 7th stadium. Under short day condi- were not incorporated into our model ab initio. The tions, in contrast, all larvae entered diapause in the developmental processes, the frequencies of dia- 5th–9th stadium at all temperatures except for pause and pupation in every simulation were in 30°C. At 30°C, some larvae (about 25%) pupated good accordance with those of the corresponding from the 4th or 5th stadium even under the short experiment. The computer program for the present day condition. It appears that P. hilaris typically model was thus verified. enters diapause unless the environmental condi- 566 Y. Watari et al.

Fig. 3. The results of simulations run under meteorological conditions at Ayabe City (34.5°N, 135.5°E), Kyoto Prefecture. Simulations were run for two or three years to estimate the development of individuals that entered diapause.

Simulations under natural conditions at Ayabe Fig. 3, the development of these overwintered non- City diapause larvae is shown assuming that they pupate Simulations under meteorological conditions at without extra molting (cf. Shintani and Ishikawa, Ayabe were conducted (Fig. 3). All individuals laid 1997c). Individuals laid on October 1 overwintered on June 1 and about 1/3 of the individuals laid on at the 1st–3rd stadium. After resuming growth the July 1 did not enter diapause and became adults next spring, about half emerged as adults in late within the year through pupation after the 4th or July–early August. Despite assumed insensitivity 5th stadium. All showed relatively concentrated to short photoperiod, the remaining larvae entered emergence in the middle of August. In contrast, the diapause at the 5th or 6th stadium in response to remaining individuals laid on July 1 and all indi- the relatively low temperature in early summer. viduals laid on August 1 entered diapause at the They passed the 2nd winter in diapause, and 5th, 6th or 7th stadium. Those individuals that en- emerged as adults in late May–early June in the 3rd tered diapause emerged synchronously in late year (Fig. 3). All individuals laid on November 1 May–early June the following year. Almost all in- overwintered in the egg stage, hatched in late dividuals laid on September 1 encountered winter March of the next year, and almost all individuals at the 5th stadium without entering diapause. In emerged as adults in late July–early August of the Life Cycle of West Japan Type P. hilaris 567 following year. The remaining individuals entered diapause at the 5th or 6th stadium and emerged in late May–early June in the 3rd year.

DISCUSSION Estimation of the life cycle of P. hilaris Simulations of the development of eggs laid on June 1–November 1 suggested that concentrated adult emergence can occur twice in one year, once in late May–early June and again in late July–early August. Since the newly emerged females are known to feed on host plant leaves for 7 to 10 d be- Fig. 4. The seasonal prevalence of west Japan type fore laying eggs (Iba, 1976), the earliest oviposi- Psacothea hilaris at Ayabe City (averages of six years between tion in one year is likely to occur from the middle 1962 and 1967). Solid line: male; broken line: female. Modi- of June. Therefore, the majority of individuals at fied from Iba et al. (1976). Ayabe are expected to enter diapause, overwinter as mature larvae, and emerge in early June of the mates information obtained by intensive laboratory following year (Fig. 3). However, some eggs laid in rearing experiments and the calculating power of a June and early July can develop to adults without computer program. Simulation of the development entering diapause and form a small emergence of west Japan type P. hilaris at any place of interest peak in the middle of August in the same year. is feasible in principle. The drawback of the model Considering the long life-span of adults (Ͼ2 is that, due to its nature, it is not readily applicable months in the laboratory), the timings of the above to other species, even to a closely-related species. two possible emergence peaks appear to corre- It is necessary to obtain detailed data on the growth spond very well with the observed adult seasonal of the species by intensive laboratory experiments. prevalence at this location (Fig. 4). Moreover, a To calculate the daylength in the field, 0.5 to portion of the individuals laid on October 1 and 1.0 h is often added to the difference between the November 1 are suggested to take two years to be- sunrise and sunset times, by taking twilight into come adults. These simulation results agree very consideration. However, in the present model, no well with those of a rearing experiment in the field such correction was made since the larvae perceive at Ayabe; 19–31% of the individuals laid on Sep- light through the bark, which is likely to reduce the tember 10–October 30 emerged as adults two years light intensity substantially. later in early June (Iba, 1976). Although this field It should be noted that the present model has experiment demonstrated the occurrence of a semi- raised questions on the overwintering of P. hilaris. voltine life cycle in P. hilaris, the mechanisms that For example, eggs laid on October 1 and November induce this phenomenon have not been discussed. 1 were simulated to overwinter as non-diapausing It appears that the west Japan type population at larvae (1st–3rd stadium) and eggs, respectively. Iba Ayabe runs basically a univoltine life cycle with a (1976) reported that not only mature larvae but lar- facultative diapause, but a portion of the popula- vae at various stadia were found from mulberry tion follows a bivoltine life cycle (partial bivoltin- trunks during winter. Therefore, in addition to dia- ism) or a semivoltine life cycle. To verify the use- pause larvae, a number of larvae and eggs (em- fulness of the present model, it is necessary to con- bryos) are likely to overwinter in quiescence, i.e., a duct simulations at various locations and compare state of non-programmed developmental arrest. Al- the results with the local prevalence of adults. Un- though eggs, neonates and pre-diapause larvae of P. fortunately, however, information on the prevalence hilaris have been shown to be highly cold hardy by of these at other locations is very scanty. laboratory experiments (Shintani and Ishikawa, 1997c, 1999), mortalities of quiescent larvae and Problems with the present model eggs during winter under field conditions have not The present model is novel in that it amalga- been carefully investigated. Thus, the significance 568 Y. Watari et al. of overwintering under a non-diapause state on the generation depending on the meteorological condi- life history of P. hilaris remains arguable. tions in May–July of that year (larger 2nd genera- In the present model, the larvae are assumed to tion population in warmer years), P. hilaris can stop development immediately after the determina- make maximum use of an unpredictable environ- tion to enter diapause is made at the midpoint of a mental resource, heat, with reduced risk for the stadium (Fig. 1). It is more likely, however, that the whole population. The life cycle of P. hilaris is in- development is arrested much later in the stadium. deed very flexible and well adapted to a changeable For more accurate estimations of the emergence environment. date of individuals that entered diapause, K and T0 for post-diapause development should be obtained Suggestions for pest management by additional experiments. Regarding the control of P. hilaris, this insect To improve the present model, it is necessary to should be controlled in the 1st to 3rd stadia, be- clarify the precise conditions for diapause induc- cause life is spent beneath the bark during these tion and termination, including the critical time stadia, while they shift into the xylem in later sta- that is sensitive to the environmental conditions. dia and thus do heavier damage to the trees (Iba, 1993). Iba (1993) found that bark application of an Significance of extra molting and partial bivol- organophosphorus insecticide, fenitrothion (MEP), tinism was effective to control the eggs and 1st to 3rd sta- In many , larger females produce dium larvae. In the present model, we can estimate more eggs but the size of each egg is fixed (Kille- the timing of adult emergence and age constitution brew and Ford, 1985; Berger, 1989; Danks, 1994). qualitatively. To establish an effective strategy to In a Cerambycid beetle, Oemona hirta, longevity control P. hilaris, however, another simulation as well as fecundity were significantly greater in model that incorporates the dynamic change in the larger females than smaller ones (Wang et al., population size is needed. Namely, we need to be 1998). Non-diapausing P. hilaris pupated from the able to predict when the populations of eggs and 4th or 5th stadium, while diapause-destined larvae 1st to 3rd stadium larvae are likely to build up. To repeated two more molts on average and entered construct a dynamic population model, it is neces- diapause mostly in the 6th or 7th stadium (Fig. 2). sary to obtain information on the pattern of ovipo- Extra molting rather than just an extension of the sition by an adult and the mortality at each devel- larval period is advantageous to increase body size, opmental stage in the field. because the larvae are freed from the limitation of ACKNOWLEDGEMENTS body size posed by the rigid exoskeleton. In fact, adults derived from diapausing larvae were signifi- We thank Professor S. Tatsuki of our laboratory for his en- cantly larger than non-diapausing counterparts couragement throughout the present study. (Shintani et al., 1996a). REFERENCES P. hilaris appears to show partial bivoltinism, as Berger, A. (1989) Egg weight, batch size and fecundity of the previously described. When heat and food are spotted stalk borer, Chilo partellus in relation to weight amply available, it is probably advantageous for the of females and time of oviposition. Entomol. Exp. Appl. species to have one more generation in the same 50: 199–207. year. However, air temperature is a highly unpre- Danks, H. V. (1994) Diversity and integration of life-cycle dictable factor. The offspring from the adults that controls in insects. In Insect Life-cycle Polymorphism (H. emerged late in the season may not have enough V. Danks ed.). Kluwer, Dordrecht, The Netherlands, pp. 5–40. time to prepare for winter, if the temperature falls Iba, M. (1976) Ecological studies on the yellow-spotted longi- sharply in autumn. Therefore, if they were to have corn beetle, Psacothea hilaris Pascoe. II. Dependence of one more generation, it would be better to become annual life cycle on the ovipositing season. J. Seric. Sci. adults from the earlier (4th–5th) stadia without Jpn. 45: 443–447 (in Japanese). extra molting, but still, to pupate and to have one Iba, M. (1980) Ecological studies on the yellow-spotted longi- corn beetle, Psacothea hilaris Pascoe. IV. The geographic more generation accompanies a risk. Partial bivol- difference of the pattern on the pronotum of adult. J. tinism appears to be a good solution for hedging Seric. Sci. Jpn. 49: 429–433 (in Japanese). against this risk. By changing the size of the 2nd Iba, M. (1993) Studies on the ecology and control of the yel- Life Cycle of West Japan Type P. hilaris 569

low-spotted longicorn beetle, Psacothea hilaris Pascoe Cerambycidae). Appl. Entomol. Zool. 27: 47–56. (Coleoptera: Cerambycidae), in the mulberry fields. Bull. Shintani, Y. and Y. Ishikawa (1997a) Temperature dependency Natl. Inst. Seric. Entomol. Sci. 8: 1–119 (in Japanese with of photoperiodic response and its geographic variation in English summary). the yellow-spotted longicorn beetle, Psacothea hilaris Iba, M., S. Inoue and M. Kikuchi (1976) Ecological studies on (Pascoe) (Coleoptera: Cerambycidae). Appl. Entomol. the yellow-spotted longicorn beetle, Psacothea hilaris Zool. 32: 347–354. Pascoe. I. The local difference in the seasonal prevalence Shintani, Y. and Y. Ishikawa (1997b) Effects of photoperiod of the adult insect. J. Seric. Sci. Jpn. 45: 156–160 (in and low temperature on diapause termination in the yel- Japanese). low-spotted longicorn beetle, Psacothea hilaris (Pascoe). Ishiishikawa, H. (1986) Control of the yellow-spotted longi- Physiol. Entomol. 22: 170–174. corn beetle, Psacothea hilaris Pascoe by the pathogenic Shintani, Y. and Y. Ishikawa (1997c) Diapause avoidance in- fungi, Beauveria tenella. Res. Rep. Ehime Seric. Exp. Stn. duced by low temperature in the yellow-spotted longicorn 8: 1–7 (in Japanese). beetle, Psacothea hilaris. Entomol. Exp. Appl. 85: 11–15. Killebrew, D. W. and N. B. Ford (1985) Reproductive tactics Shintani, Y. and Y. Ishikawa (1998) Photoperiodic control of and female body size in the green lynx spider, Peucetia larval diapause in the yellow-spotted longicorn beetle, viridans (Araneae, Oxyopidae). J. Arachnol. 13: 375– Psacothea hilaris (Pascoe): Analysis by photoperiod ma- 382. nipulation. Entomol. Exp. Appl. 86: 41–48. Kusama, K. and M. Takakuwa (1984) . In Longicorn- Shintani, Y. and Y. Ishikawa (1999) Geographic variation in Beetles of Japan in Color (Japanese Society of cold hardiness of eggs and neonate larvae of the yellow- Coleopterology ed). Kodansya, Tokyo, pp. 374–493 (in spotted longicorn beetle Psacothea hilaris. Physiol. Ento- Japanese). mol. 24: 157–164. Lindsey, A. A. and J. E. Newman (1956) Use of official Shintani, Y., Y. Ishikawa and S. Tatsuki (1996a) Larval dia- weather data in spring time-temperature analysis of an In- pause in the yellow-spotted longicorn beetle, Psacothea diana phonological record. Ecology 37: 812–813. hilaris (Pascoe) (Coleoptera: Cerambycidae). Appl. Ento- Makihara, H. (1986) Longicorn beetle. Expansion of the dis- mol. Zool. 31: 489–494. tribution and geographic variation. In Japanese Insect. Shintani, Y., S. Tatsuki and Y. Ishikawa (1996b) Geographic Ecology of Invasion and Disturbance (K. Kiritani ed.). variation of photoperiodic response in larval development Tokai Univ. Press, Tokyo, pp. 96–106 (in Japanese). of the yellow-spotted longicorn beetle, Psacothea hilaris Roltsch, W. J., F. G. Zalom, A. J. Strawn, J. F. Strand and M. J. (Pascoe) (Coleoptera: Cerambycidae). Appl. Entomol. Pitcairn (1999) Evaluation of several degree-days estima- Zool. 31: 495–504. tion methods in California climates. Int. J. Biometeorol. Suzuki, E. and T. Yoshii (1980) Seasonal prevalence and varia- 42: 169–176. tion of spot pattern of adults of the yellow-spotted longi- Sakakibara, M. (1995) Egg periods in several populations of corn beetle, Psacothea hilaris (Pascoe). Rep. Tohoku yellow-spotted longicorn beetle, Psacothea hilaris Pascoe Seric. Exp. Stn. Abst., p. 32 (in Japanese). (Coleoptera: Cerambycidae). Jpn. J. Appl. Entomol. Zool. Wang, Q., G. L. Shi and L. K. Davis (1998) Reproductive po- 39: 59–64 (in Japanese with English summary). tential and daily reproductive rhythms of Oemona hirta Sakakibara, M. and K. Kawakami (1992) Larval diapause in- (Coleoptera: Cerambycidae). J. Econ. Entomol. 91: heritance mode in two ecotypes of the yellow-spotted 1360–1365. longicorn beetle, Psacothea hilaris (Pascoe) (Coleoptera: