Developmental and reproductive response of Brachmia macroscopa (Lepidoptera: Gelechiidae) to three host plants

The sweet potato leaf folder, Brachmia macroscopa Meyrick (Lepidoptera: Gelechiidae), which is a significant pest of plants in the family Convolvulaceae, is rapidly expanding its range in South China and other subtropical regions. Studies were designed to examine the effects of three different host plants (sweet potato, Ipomoea batatas (L.) Lam.; water spinach, I. aquatica Forsskål; and morning glory, Pharbitis purpurea (L.)) on the development and life table parameters of B. macroscopa under laboratory conditions. We found that the intrinsic rates of increase of B. macroscopa were 0.17 ± 0.004, 0.21 ± 0.005 and 0.16 ± 0.004 on I. batatas, I. aquatica and P. purpurea, respectively. The highest net reproduction rate was 158.06 ± 18.22 per female reared on I. aquatica. The larvae had five instars when reared on I. batatas and I. aquatica, but required six instars on P. purpurea. The mean generation lengths of B. macroscopa ranged from 24.32 ± 0.18 days when reared on I. aquatica to 29.40 ± 0.24 days on P. purpurea. The survival of all stage and fecundity curves was intuitively manipulated using the age-stage-structured and two-sex population life table method, to enable comprehensive descriptions of the stage and population trends of B. macroscopa on the three Convolvulaceae plants. Our results indicated that I. batatas and I. aquatica were more suitable host plants than P. purpurea.

The relationships between insects and their host plants are considered to be co-evolutionary interactions. In the process of long-term evolution and adaptation to their environment, insects have gradually produced selectivity to their hosts. The host preference may not only influence the growth and development of insects, it may also play an important role in the growth of their population 1 . Feeding on different host plants can also affect the susceptibility of some insect species to certain pesticides 2 . Some studies have shown that the response of insects to the host is based on their demand for external nutrition, while other researchers are of the opinion that the insects are affected by secondary metabolites in the host plants 3,4 . In general, when insects fed on plants that they are adapted to, their growth and survival rates will be optimal and their reproductive capacity will be at a high level, and vice versa 5 . Host plants, in order to discourage insect feeding behavior, are also continually developing an array of defensive strategies, including phenological defenses 6,7 , physical defenses 8 , chemical defenses 9 , etc. The overuse of chemical pesticides to control lepidopteran (and other) pests during the past few decades, has forced targeted insects to evolve in terms of inherent behavioral mechanisms, and altered ecological, genetic and physiological factors, with the end result being the development of resistance to many chemical controls 10 . For example, in using the insecticide fenvalerate in the field to control the diamondback moth, Plutella xylostella (L.) (Lepidoptera: Plutellidae), and spraying 10 or more times a year for only two years, has resulted in local populations of the moth that are as much as 12,000 times more resistant to pyrethroid insecticides than other populations 11 .
The sweet potato leaf folder, Brachmia macroscopa Meyrick (Lepidoptera: Gelechiidae), is a widely distributed pest in Europe, Russia, Caucasus, Transcaucasian region, West Kazakhstan, central Asia, Korea, Japan, China, and northern India. It mainly damages Dioscorea esculenta, Ipomoea aquatica, Calystegia sepium, Croomia japonica and other crops in the Convolvulaceae family. The damage to crops by larvae is significant greater than it is the adult stage. The host plant leaves are damaged by the larvae skeletonizing the surface, leaving only the transparent leaf epidermis; the mesophyll layer of the leaf is completely consumed. The harm caused by B. macroscopa larvae seriously affects the ability of the crops to photosynthesize, causing the host plant leaves to wither, and eventually die. The larvae frequently cause extensive damage to their host crops and result in serious economic loss to the growers. Our previous studies have involved determining the complete mitochondrial genomes of this pest, as well as detailed research on the developmental and fecundity responses of the species under different temperatures to identify the most susceptible phase of its development and growth 12,13 . Additionally, based on a two-year field investigation in Changsha, China, we found that local B. macroscopa fed mainly on three Convolvulaceae plants, sweet potato vine Ipomoea batatas (L.) Lam., Chinese waterspinach I. aquatica Forsskål, and Japanese morning glory Pharbitis purpurea (L.) 14 . In the present study, we determined different life parameters of B. macroscopa by constructing life tables for the moth when reared on each of the three major hosts. The data provides a theoretical basis for devising and implementing efficient as well as environmentally friendly integrated control procedures.

Results
Larval stages. B. macroscopa was capable of completing its entire life cycle when fed on each of the three host plants tested, although the duration of each stage varied ( Table 1). The developmental time of the egg stage was 4 days on I. aquatica and P. purpurea, which was shorter than eggs laid on I. batatas (5 days). The B. macroscopa larval developmental durations were significantly different (P < 0.0001) when reared on the three plants. When fed on P. purpurea, the larvae required six instars to complete development with a developmental period of 14.36 ± 0.14 d, which was the longest total growth period among the three host species. The larval development on I. batatas and I. aquatica required only five instars, and lasted 12.42 ± 0.11 d and 11.50 ± 011 d, respectively. Significant differences occurred in the developmental duration (P < 0.0001) of the 1st and 5th larval stage when reared on the three host plants, but were not found in the 4th instar duration (P = 0.536). There was no significant difference between I. batatas and I. aquatica in the duration of the 2nd instar (P = 0.448), or in the 3th instar (P = 0.116); nor were there significant differences in the 3th instar stage between I. aquatica and P. purpurea (P = 0.454), although the duration of the 2nd instar when fed on I. aquatica was longer than on P. purpurea. No significant difference occurred in the 3th instar between I. batatas and I. aquatica (P = 0.116). The 2nd instar larval development duration when fed on I. aquatica was longer than on P. purpurea. The reverse, however, was true during the 3rd and 4th instars. There were no significant differences between the lengths of the pupal periods after being fed on I. batatas and I. aquatica (P = 0.479). The total durations of the larval developmental periods of B. macroscopa fed on the three host plants did show significant differences: the relationship of the lengths was: P. purpurea > I. batatas > I. aquatica, (24.06 ± 0.18 d, 22.50 ± 0.12 d and 20.48 ± 0.10 d, respectively). Oviposition period and fecundity. Significant differences (P < 0.05) were found in B. macroscopa adult females reared on the three host plants for the adult pre-oviposition period (APOP; the period of time between the emergence of an adult female and the onset of her first oviposition), the total pre-oviposition period (TPOP; the time interval from birth to the beginning of oviposition), the oviposition period and total fecundity (total number of eggs produced during an individual's reproductive period) ( Table 3). Although the APOP of B. macroscopa female adults from I. batatas (1.45 ± 0.12 d) and P. purpurea (1.39 ± 0.10 d) were similar and not significantly different (P = 0.734), the APOP on I. aquatica (0.44 ± 0.10 d) was significantly shorter (P < 0.05) than the other two groups. The results demonstrated that the fecundity of adult females from I. aquatica (427.20 ± 18.08) was significant higher (P < 0.05) than it was in the I. batatas (298.18 ± 11.42) and P. purpurea (299.03 ± 8.86) groups.

Total duration and longevity of the adult stage. It is evident from
Since only female adults produce offspring, there is only a single curve representing the female age-stage-specific fecundity (f x8 ). The peak age-stage-specific fecundity (f x8 ) on the three different host plants occurred at  Life expectancy and reproduction rate contribution value. The age-stage life expectancy (e xj ) is the expected lifespan of different individuals at age x and stage j when fed on different host plants (Fig. 3). Brachmia macroscopa had the longest life expectancy on I. batatas, followed by I. aquatica and then P. purpurea. The results showed that B. macroscopa growth on I. batatas was relatively slow. The reproductive value (v xj ) is the contribution of different individuals at age x and stage j to future population growth (Fig. 4). The reproductive value reached its highest peak on the 24th day (V 24    (P > 0.05) higher r and λ values than the other two groups, although the T on I. aquatica (24.32 ± 0.18 d) was shorter than it was on I. batatas (28.76 ± 0.23 d) and P. purpurea (29.40 ± 0.24 d) (P < 0.05).

Discussion
The development period and fecundity of lepidopterous insects, as well as insect populations in general, are influenced by many factors, including the availability of suitable host plants. The selection of appropriate local host plants plays a vital role in determining whether a natural population will grow or decline. An ideal means of deciphering the factors that enter into this selection is via the use of life tables. Life tables are also useful in providing applied and practical information regarding a given pest's survival, fecundity and developmental times. Moreover, difference in the individual's growth and developmental when feeding on different plants can also be determined. Damage caused by B. macroscopa to the three widespread hosts selected for this study, I. batatas, I. aquatica and P. purpurea, has been widely recorded in many countries besides China 15 . Because insecticides have traditionally been used to suppress outbreaks of this pest 16 , it is hoped that further understanding the life history of B. macroscopa will allow growers to decrease their dependance on pesticides in attempts to control this widespread pest in the future.
In the present study we used the age-stage life table to compare the effects of the above three host plants on the life history parameters of B. macroscopa. The age-stage-specific survival rate (S xj ) indicates the survival rate at different stages as well as demonstrating the generation overlapping phenomenon. The study of the life table of the geometrid Problepsis superans (Butler) on different hosts indicated that at the age of 13 d, the 1st instar larvae coexisted with the 2nd and 3rd instars when P. superans fed on Ligustrum lucidum Aiton, while the insect popu-  18,19 . This shows that the overlapping generations that occur in insect life histories should not be overlooked. While this phenomenon is an integral component of the age-stage life table survival curve, this critical information is lacking in traditional life tables. In the present study, it is evident from Fig. 3 that B. macroscopa obviously has this overlapping phenomenon in the development of its immature stages and can exist in multiple stages on any given day. The egg stage lasted 4 days on I. batatas and I. aquatica, and 5 days on P. purpurea. This was within the range (3-6.35 days) reported by Ma et al. when studying the effects of varying temperature on B. macroscopa 13 . The length of the egg stage was also close to that observed in several other species of Gelechiidae. A study by Karimi-Pormehr et al. 20 reported that the length of the egg stage of, Sitotroga cerealella (Olivier) (Lepidoptera: Gelechiidae) on ten different barley cultivars was from 4.83 to 4.97 days. Rostami et al. 21 reported the egg stage of Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae) lasted from 4.43 to 4.64 days on three tomato cultivars, but other researchers, found the egg stage ranged from 5 to 7 days in the species 22,23 , B. macroscopa also has a shorter pupal stage compared to T. absoluta. In our study, the B. macroscopa pupal stage ranged from 5 to 6 days,  20 . In T. absoluta and C. ohridella, the fecundities were consistently less than 200 eggs [21][22][23][24] . This high fecundity may explain the multiple outbreaks that are increasingly commonplace in recent years, causing serious damage to a number of crops. Developmental duration and fecundity of the insects were measured to determine whether the herbivorous insects had a good "biological fit" on a particular host plant or not. A short developmental duration and high fecundity may optimize a more suitable environment for the insects 25 . When B. macroscopa was reared on I. aquatica, the total duration of the egg and larval stages were the shortest, but the survival rates in the larval stage, and the adult fecundity were the highest, suggesting that I. aquatica may be the most suitable host for B. macroscopa. When evaluating the biological characteristics of an insect population, it is important to not only include the growth and developmental duration, feeding and oviposition abilities 26 , while the r and R 0 data, as well as other population dynamics should also be considered [27][28][29] .
The intrinsic rate (r) and net reproductive rate (R 0 ) of B. macroscopa was 0.16 to 0.21 days −1 and 122.01 to 158.06 eggs, respectively. These numbers were higher than in some other gelechiid species. Karimi-Pormehr et al. 20 reported that the r and R 0 values for S. cerealella were 0.077 to 0.115 days −1 and 17.17 to 44.09 eggs, respectively depending on the barley cultivar. The r and R 0 values for T. absoluta were 0.074 to 0.115 days −1 and 10.03 to 61.56 eggs, respectively, on different tomato cultivars 21,22 . The r of B. macroscopa was similar to several noctuid species, e.g., in Helicoverpa armigera it was 0.115 to 0.142 days −1 and 0.171 to 0.264 days −1 in Spodoptera exigua (Hübner) 19,30 , but the R 0 values were less than those in H. armigera (177.3 to 270.1 eggs) and S. exigua (126.4 to 377.1 eggs) 17,30 . The r value of B. macroscopa was smaller than that reported for P. xylostella which exceeded 0.2 days −1 31,32 . The R 0 of P. xylostella differed significantly on various host plants ranging from 30.6 to 183.8 eggs 31,32 . The R 0 of B. macroscopa, under an ideal temperature of 27 °C, was reported as 147.60 eggs by Ma et al. 13 . Our results approximated theirs (122.01 to 158.06 eggs). The B. macroscopa population appeared to be especially stable, since its host selectivity was not especially strong.
The primary reason for animals (including insect herbivores) to eat is to acquire a mix of nutrients in order to complete the processes of growth, development, and reproduction 33 . On the contrary, different host plants can also affect insects' growth and reproduction, which can potentially lead to a change of their life history 34,35 . These changes are often vital to the insect's population dynamics in the wild. The nutrition contained in a host plant determines whether it suitable for an insect population or not. Of prime importance, the quality and quantity of the plants can affect the food digestibility and fecundity of a pest; secondarily, a high fecundity and short TPOP will directly lead to a high intrinsic rate 19 . When B. macroscopa larvae fed on I. aquatica, the total fecundity was higher than in the other two hosts with 427.60 eggs, the TPOP was the shortest with 20.58 days and the intrinsic rate (r) was also the highest, as were the R 0 , GRR and λ values. These results would suggest that I. aquatica is the most suitable host for B. macroscopa. The r and λ values of B. macroscopa were the second highest on I. batatas. Insects on I. batatas had the shortest pupal period as well as the longest adult duration. These results indicated that B. macroscopa was able to develop adequately on I. batatas, did not do well on P. purpurea. Finally, B. macroscopa apparently has considerably stronger host adaptability on Ipomoea than it does on Pharbitis plants.

Materials and Methods
Tested plants. Based    host to encourage mating and to act as an oviposition substrate. A mesh net and a cotton ball soaked with a 20% honey solution were placed in the rearing cages for ventilation and nutrition, respectively. The insects used in the experiments had been laboratory bred for at least two generations on each host species. All tested insects were maintained in an artificial climate chamber at 27 ± 1 °C, 75 ± 20% RH and 10 L: 14D, and reared on one of the three host plants mentioned above.
Experiments. One hundred and fifty healthy, well developed, recently laid eggs that were oviposited during the spawning peak, were collected and placed in an artificial climate chamber set at 27 ± 1 °C, 75 ± 20% RH, and a photoperiod of 10:14 (L: D) h. Hatching was observed and recorded daily. The newly hatched larvae were transferred using a brush into petri dishes (15 cm in diameter, 1.5 cm in height) and provided with fresh leaves from I. batatas, I. aquatica or P. purpurea plants. The petioles of these leaves were inserted into water-soaked cotton to maintain freshness. The larval mortality was recorded and the leaves replaced daily. Newly pupated individuals were transferred into separate glass tubes (2 cm in diameter, 10 cm in height) with a water-soaked cotton ball to retain moisture and covered with gauze. The pupae were observed daily until emergence as adults. Detailed records were kept on the durations and survival of each larval instar, pupal stage and adult longevity. Upon eclosion, females and males emerging within a 24 h period from each of the host plants were paired and maintained in a plastic oviposition chamber (13 cm in diameter, 17 cm in height). A small host plant, planted in a disposable cup, was added to each container as an oviposition substrate. A cotton ball soaked with a 20% honey solution was placed in the cup to serve as a food source for adults and was replaced daily. The numbers of newly produced eggs were totaled every 24 h until all the adults had died.
Life table parameters. The age-stage, two-sex life table method was used to analyze the raw data. The age-stage-specific survival rates (s xj ) were determined for each of the three treatments. These are the survival rates at age x (days) and stage j (where the first through eighth stages represent the egg stage (1), the 1st through 4th larval instars (2-5), the 5th (or combined 5th + 6th -when present) instars (6), the pupal stage (7) and the adult stage (8), respectively) 36 . The age-stage specific fecundity (f xj ) is defined as the quantity of eggs produced by each individual per day at age x (days) and stage j; the age-specific survival rate (l x ) is the probability that a newly produced egg will survive to age x; the age-specific fecundity (m x ) is the mean number of female eggs laid per female adult at age x. The population parameters were calculated according to the procedure developed by previous researchers 36 . The intrinsic rate of increase (r) was calculated using the iterative bisection method where . The life parameters GRR (total reproductive rate), λ (finite rate of increase) and T (mean generation time, which is the duration that a population needs to increase to R 0 -fold its size before reaching the stable age-stage) were calculated as GRR = λ ∑ = m e , x r m and T = (lnR 0 )/r m . Finally, the means, standard errors and variances of the population parameters were bootstrapped using the computer program TWOSEX-MSChart 38 . The graphs were created using Sigma plot 12.5.

Data analysis.
The raw life-history data of B. macroscopa reared on each host species were entered separately into Microsoft Excel 2015, and the data analyzed by one-way ANOVA using SPSS 22.0. The correlation coefficient differences were compared using the Tukey method (P < 0.05).