Assessment of Trichogramma japonicum and T. chilonis as Potential Biological Control Agents of Yellow Stem Borer in Rice

Two species of Trichogramma wasps were assessed for their effectiveness against yellow stem borer Scirpophaga incertulas. A laboratory cage test with T. japonicum and T. chilonis showed that both species parasitized yellow stem borer egg masses at 60.0% ± 9.13% and 40.7% ± 7.11%, respectively, with egg parasitism rates of 15.8% ± 22.2% for T. japonicum and 2.8% ± 5.0% for T. chilonis. Once the host eggs were parasitized, emergence rates were high for both species (95.7% ± 0.12% for T. japonicum and 100% for T. chilonis). In paddy field trials, the two Trichogramma species were released at three densities (50,000/ha, 100,000/ha and 200,000/ha) in Southwestern China. Egg mass parasitism was 9% ± 7.7% for T. japonicum and 15% ± 14.1% for T. chilonis, and again only a relatively small fraction of eggs was successfully parasitized. No clear conclusion could be drawn on the most efficient release rate as no significant differences were found among the three release rates. A comparison of field-collected T. japonicum with T. japonicum and T. chilonis mass reared on Corcyra cephalonica showed significantly larger body size and ovipositor length in field-collected wasps, suggesting potentially higher effectiveness on yellow stem borer eggs after at least one generation on the target host. Factors contributing to the low field parasitism rates are discussed.


Introduction
Rice (Oryza sativa L.) is the most important crop in the world [1] and plays a central part in Asian food security [2]. This crop is widely distributed especially in Southern parts of China [3] where rice accounts for 88.7% of the total agricultural acreage [3]. Yunnan Province is located in the Southwest of China, and is considered to be part of the Greater Mekong Subregion, together with Myanmar, Laos, Cambodia, Vietnam and Thailand [4]. Rice covers over half of the cropping lands in this region [5] and is considered to be the most important crop there [5,6]. However, rice production in this area is suffering from serious pest and disease damage [7][8][9][10] causing substantial yield losses every year [6,[11][12][13]. Yellow stem borer moths were collected from nearby rice fields by sweep nets and then caged for oviposition on rice plants. Egg masses were collected daily from the plants and used within 24 h in the tests. Rice plants at heading stage randomly selected from the paddy fields near the experimental site were transplanted together with soil to pots and used for YSB oviposition and in cage tests.

Performance of the Trichogramma Species on Substitute Host
A test was conducted to measure the performance of both wasp species on the substitute host S. cerealella. Five hundred fresh, UV-sterilized Sitotroga eggs were each exposed to 100 T. japonicum and T. chilonis wasps in a Petri dish. After 48 h, the wasps were removed and the eggs were transferred to a Petri dish lined with moistened filter paper under room temperature. The emerged Trichogramma adults were counted and sexed daily. The measurement was replicated three times for each species.

Cage Tests for Parasitism
Cage tests were conducted to assess the capacity of T. japonicum and T. chilonis to parasitize YSB eggs. Rice hills at early heading stage were collected from the field and grown in pots in cylindrical cages (120 cm in height and 35 cm in diameter) made of wire and fine netting at natural conditions (1 to 2 hills per cage). Twenty pairs of yellow stem borer adults were introduced into a cage for oviposition. After 24 h, 1/5 of a tricho-card of either T. japonicum or T. chilonis that was ready to emerge was introduced per cage, corresponding to an estimated 100 parasitized eggs for each of the two species tested. Four replicates (cages) were set up for each Trichogramma species. Forty-eight hours after introduction of tricho-cards, all YSB egg masses were cut down together with rice leaves and placed individually on a moistened filter paper in glass tubes (15 cm in length and 2.5 cm in diameter) under room temperature. The egg masses were observed daily for hatching of yellow stem borer larvae, if any (and hatched larvae were removed), or emergence of Trichogramma wasps. Number of hatched larvae and emerged wasps were recorded. The emerged wasps were sorted by sex. Number of dead unparasitized eggs and dead parasitized eggs were differentiated by dissection under a stereo microscope after wasp emergence had ceased for three days. Egg mass parasitism rate was calculated by dividing numbers of parasitized egg masses with total numbers of egg masses. Egg parasitism rate was calculated by dividing numbers of parasitized eggs with total numbers of eggs.

Field Tests for Parasitism
Field performance of the two Trichogramma species was tested by releasing the wasps in paddy fields. The fields were planted to conventional rice varieties of "Deyou 8", "Deyou 12" and "Deyou 16" that were at heading to milking stage when the experiments were conducted. Insecticides were not applied for one week before releasing of Trichogramma and during the experiment period.
Both T. japonicum and T. chilonis were released at 100 points per ha at 500, 1000 and 2000 wasps per point, corresponding to release densities of 50,000, 100,000 and 200,000 wasps/ha, respectively. For each combination of release density and Trichogramma species, there were four replicates (field plots). Another four plots were used as control. Each plot covered an area of 900 m 2 (30 m × 30 m), and was at least 15 m (edge to edge) away from the neighboring plots. Tricho-cards were released in a plot at nine points that were 10 m apart from each other. The plots were randomly assigned to treatments in the field (Appendix A Figure A1).
On the same day when the blackened (ready-to-emerge) tricho-cards were released, sentinel YSB egg masses less than 24 h old were placed in the plots. The sentinel egg masses were obtained from the established cage rearing. For each plot, 10 YSB egg masses together with a rice leaf segment were randomly fixed on either leaf side by a stapler. Forty-eight hours after releasing, YSB egg masses were recollected from the fields and taken back to laboratory. In addition to re-collecting the actively placed egg masses, all other YSB egg masses found were also taken back to the lab. Egg masses were kept under natural conditions and observed for number of hatched larvae, emerged wasps and dead eggs, as described in the cage tests. The local weather in July and August of 2013 was characterized by small rains occurring at least once a day, with a temperature range of 22 • C to 33 • C and an average wind speed less than 3.4 m/s.

Morphological Observations
Morphological parameters potentially affecting the ability to parasitize YSB eggs were measured for field-collected and mass reared Trichogramma. The field-collected T. japonicum were from YSB egg masses collected in 2014 from paddy fields in Husa Township where T. japonicum had been released for one year. The mass reared T. japonicum and T. chilonis were produced by the local Trichogramma rearing facility in Mangshi, Dehong Prefecture, where the wasps had been reared for >20 generations on eggs of C. cephalonica. Body length, ovipositor length and hind tibia length were measured under a stereo microscope to 15 wasps for each of the field-collected T. japonicum and the mass reared T. japonicum and T. chilonis.

Data Analysis
Egg parasitism rate was calculated as: (number of dead parasitized eggs + number of emerged wasps)/(total number of eggs); egg mass parasitism rate, as number of parasitized egg masses/total number of egg masses; and emergence rate, as number of emerged wasps/(number of dead parasitized eggs + number of emerged wasps). For the field test, mean values of the parameters were calculated for a plot and then tested with a two-factorial ANOVA for the two factors "species" and "densities". For the cage test, means of the parameters between the two species were differentiated by t-test or Mann-Whitney U-test. All proportional data were arcsin-squareroot transformed before analyses, which were performed with the program IBM SPSS Statistics 19.0.0 (SPSS Inc., Chicago, IL, USA).

Performance of the Trichogramma Species on Substitute Host
Ninety percent of the Sitotroga eggs were parasitized in the test. From a total of 500 eggs exposed for each of the two Trichogramma species, there were 289 ± 46 T. japonicum wasps and 245 ± 18 T. chilonis wasps that emerged, respectively. No significant difference was found in mean emergence rates between T. japonicum (64.2% ± 10.2%) and T. chilonis (54.4% ± 4%; t = 0.902, p = 0.42).

Cage Tests
A total of 77 YSB egg masses were collected from the cage tests, 37 for the T. chilonis treatment and 40 for the T. japonicum treatment. Mean parasitism rates of egg masses were 60.0% ± 9.1% for T. japonicum and 40.7% ± 7.1% for T. chilonis and did not significantly differ from each other (t 6 = 1.67, p = 0.145; Figure 1a).

Cage Tests
A total of 77 YSB egg masses were collected from the cage tests, 37 for the T. chilonis treatment and 40 for the T. japonicum treatment. Mean parasitism rates of egg masses were 60.0% ± 9.1% for T. japonicum and 40.7% ± 7.1% for T. chilonis and did not significantly differ from each other (t6 = 1.67, p = 0.145; Figure 1a).

Field Tests
In total, 370 YSB egg masses were recollected from the 28 plots of which 280 were sentinel egg masses and 90 were naturally laid. This included 43 egg masses from control plots, 165 from T. japonicum released plots and 162 from T. chilonis released plots. Because no significant difference was observed between sentinel eggs and naturally laid eggs in egg parasitism rates (Appendix A Figure A2), data from all egg masses were pooled for further analysis. For the 43 egg masses collected from control plots, no attack by Trichogramma was observed, resulting in zero parasitism rate. In the Trichogramma release plots, parasitism rates of YSB egg masses were 9.0% ± 7.6% and 15.1% ± 14.1% for T. japonicum and T. chilonis, respectively. Rather low parasitism rates of YSB eggs were found in the present study with 0.35% ± 0.36% and 0.68% ± 0.66% for T. japonicum and T. chilonis, respectively. No significant differences were observed between the two Trichogramma species, neither for egg mass parasitism rate (excluding the control plots showing zero parasitism, ANOVA, F 1,18 = 0.84, p = 0.37) or for egg parasitism rate (F 1,18 = 1.20, p = 0.29, Figure 2). The emergence rates were 84.4% ± 19% for T. japonicum and 83.4% ± 15% for T. chilonis (F 1,18 = 1.7, p = 0.2). Furthermore, no significant differences were found among the three release densities tested, neither for egg mass parasitism (F 2,18 = 0.42; p = 0.66) nor for egg parasitism (F 2,18 = 1.70 p = 0.21, Figure 2). A significant interaction effect was observed between wasp species and release density for egg mass parasitism rate (F 2,18 = 4.14, p = 0.033) but not just for egg parasitism rate (F 2,18 = 3.21, p = 0.064). For cage tests, a weak but significant negative correlation was found between egg mass size (indicated by total egg number of single egg mass) and egg parasitism rate (Appendix A Figure A3, Pearson correlation: R 2 = −0.349, p = 0.002). Yet no significant correlation could be observed between egg mass size and egg parasitism rate in field tests (Pearson correlation: between wasp species and release density for egg mass parasitism rate (F2,18 = 4.14, p = 0.033) but not just for egg parasitism rate (F2,18 = 3.21, p = 0.064). For cage tests, a weak but significant negative correlation was found between egg mass size (indicated by total egg number of single egg mass) and egg parasitism rate (Appendix Figure A3, Pearson correlation: R 2 = −0.349, p = 0.002). Yet no significant correlation could be observed between egg mass size and egg parasitism rate in field tests (Pearson correlation: R 2 = −0.149, p = 0.451).

Morphological Comparison
The field-collected Trichogramma wasps were identified as T. japonicum (Appendix Figure A4). Significant differences between field-collected T. japonicum and mass reared T. japonicum or T. chilonis were found for body length (ANOVA, F2,36 = 93.5, p < 0.001), ovipositor length (F2,36 = 51.4, p < 0.001) and hind tibia length (F2,36 = 5.88, p = 0.0062, Figure 3). In particular, the field-collected T. japonicum were larger than the mass reared conspecifics (Tukey HSD test, body length: p < 0.001, ovipositor length: p < 0.001), although no significant difference was found for hind tibia length (p = 0.134). Furthermore, the mass reared T. japonicum were larger than the mass reared T. chilonis for two tested parameters (body length: p < 0.001, ovipositor length: p < 0.001, hind tibia length: p = 0.242) Figure 3. Comparison of body length, ovipositor length and hind tibia length among Trchogramma species/strains either collected from yellow stem borer egg masses from paddy fields (wild T. japonicum) or reared in Trichogramma rearing facilities (TRFs) on eggs of C. cephalonica (n = 15). Error bars indicate + SE. Lower case letters indicate significant differences among the three species/strains (one-way ANOVA followed by Tukey HSD multiple comparison).

Discussion
In the present study, we tested two Trichogramma species collected in the target region for their potential as a biological control agent of YSB, the main pest of rice in the target region. No conclusion can be drawn as to which of the two tested species might be better, since results from cage and field tests were not consistent and slightly contradictory. Overall, parasitism rates obtained for T. japonicum and T. chilonis in both the cage and field tests were relatively low and raise concerns as to

Discussion
In the present study, we tested two Trichogramma species collected in the target region for their potential as a biological control agent of YSB, the main pest of rice in the target region. No conclusion can be drawn as to which of the two tested species might be better, since results from cage and field tests were not consistent and slightly contradictory. Overall, parasitism rates obtained for T. japonicum and T. chilonis in both the cage and field tests were relatively low and raise concerns as to whether these species could be successfully applied. However, a number of reasons may have contributed to low egg mass parasitism rates in the field. First of all, it cannot be ruled out that the quality of the released material is impaired, which may be due to being mass produced for many generations on Sitotroga [26,29], or due to sub-optimal conditions during transportation of the tricho-cards from the production facility in Hengshui to the release site in Husa Township. The low emergence rates found in the substitute host tests on Sitotroga eggs in the present study provide some evidence that quality of the wasps used in the study here was below optimum. It is furthermore well known for field release studies with inundative releases in plots that dispersal from these plots can be an issue and this has also been shown for Trichogramma [30]. Even though the release area was a bit larger than the area from which measurements were taken, dispersal over 10-20 m can happen for Trichogramma with corresponding effects on parasitism. In addition, the experiment was run over a short time only to avoid too much loss of egg masses due to predation. It is also likely that not all Trichogramma emerged during the first day of the experiment, i.e., fewer Trichogramma wasps may have been active during the length of the experiment than would be anticipated from the release rate. Last but not least, Trichogramma wasps may have had difficulties to find egg masses because they were placed experimentally on leaves which may quickly dry in the field and thus become unattractive. In light of the above considerations and compared to other studies on similar scales, the relatively low parasitism rates may be partly explained; rates would perhaps be higher in larger scale releases. Also, the fact that this study demonstrated no consistent effect of the release density indicates that unknown factors may have played a role in this field experiment.
However, despite the fact that egg mass parasitism may have been underestimated in the present study, a particular concern is still the low parasitism rates that were found for eggs, i.e., even though an egg mass might have been parasitized, Trichogramma females only parasitized a small proportion of that egg mass. This may be related to the specific features of YSB egg masses which on one hand consist of several layers of eggs, and on the other hand are also covered with hairs provided by the YSB female moth. In general, few studies have been carried out to test Trichogramma wasps on yellow stem borers and in particular, field surveys have mostly only reported egg mass parasitism rates instead of egg parasitism rates for Trichogramma wasps toward YSB. For egg masses, exceptional rates of 100% parasitism could be reached for T. japonicum toward the first generation of YSB eggs in the fields [11]. In contrast, several surveys have reported that parasitism rates of YSB eggs by Trichogramma in paddy fields ranged only from 2.1% to 23.1% [31][32][33]. These rates are considerably lower than those generally found for striped stem borer eggs, where parasitism rates reach up to 85.8% [6,21].
Hairs of pests have been proved to be related to both direct [34] and indirect [35] protection from natural enemy attacks. Similarly, due to the special hair cover structure and layers of eggs inside, YSB eggs usually suffer less from egg parasitoid wasps [25]. This is supported by Lou et al. [11] who recently showed that the eggs from deeper layers of YSB egg masses usually escape from being parasitized by Trichogramma wasps. At the high densities used for the field cage study here (100 wasps per 10 egg masses), it may be concluded that wasps would parasitize all eggs they could possibly reach which would suggest about 21% for T. japonicum and 4% for T. chilonis on small egg masses and 4.7% for T. japonicum and 1.9% for T. chilonis on larger egg masses. In fact, a negative correlation of egg parasitism rates with egg mass size was observed in cage tests, which indicates that small egg masses were more likely to be parasitized in a higher proportion than large egg masses. No such correlation was found in the field test. We were not able in the present study to precisely analyse the rate of parasitization of eggs in the different layers of egg masses, and additional studies would be worthwhile to deepen our understanding on the factors underlying our findings. Even though the hairs on the egg surface are likely impairing parasitism rates, Trichogramma offspring may be expected to benefit from hairs of parasitized YSB egg masses [36] as predation risk is reduced compared to Trichogramma offspring inside striped stem borer or rice leaf folder eggs [34,35] thus increasing offspring survival rate [37].
In the experimenta, no Trichogramma at all were found in the control fields or during additional collections in Husa Town [38]. Although the reasons for this remain unknown, it may be speculated that the high rate of recent pesticide applications is a contributing factor. Interestingly, Trichogramma wasps were recovered two months after the last of six releases that were conducted in 20 hectares of demonstration paddy fields located in Mangshi, Dehong, Yunnan Province, indicating that a natural T. japonicum population has successfully established there. It is known that body size of host parasitoid wasps can be determined by host egg sizes [39,40], and we found that field-collected T. japonicum strains were significantly larger than lab strains, produced on rather small eggs of the factitious host Sitotroga, and also had longer ovipositors. Larger body sizes of parasitoids are generally related to higher egg load, lifetime fecundity, ability to disperse, host locating ability, and thus may contribute to increased parasitism rates [39][40][41]. This suggests that Trichogramma wasps reared in small eggs might encounter difficulties when facing larger eggs or packs of egg masses, especially when they are covered with hairs in the field. In this case, second generation wasps emerging from field hosts may be more successful in parasitizing YSB eggs.

Conclusions
In conclusion, the overall low parasitism rates found in the present study suggest that the two Trichogramma species tested would not be highly successful for inundative biological control of YSB, in particular because of the low attack of eggs observed. However, considering the limitations of such an experimental study (see points discussed above) and likely higher success rates of larger wasps emerging from the field host, application of Trichogramma could have a positive effect on YSB pest control based on inundative or even inoculative releases in the longer term, and further studies will be needed to fully understand the potential, but also constraints, of the present system.

Conflicts of Interest:
The authors declare no conflicts of interest.      Figure A4. The Trichogramma evolutionary history was inferred using the Neighbor-Joining method [42]. The optimal tree with the sum of branch length = 0.94808551 is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches [43]. Scale indicates evolutionary distances among species. A total of nine Trichogramma species and yellow stem borer were picked up for phylogenetic analysis. Ten cytochrome oxidase subunit I genes were introduced from gene bank for the analysis including T. japonicum (gi|727353589|), T. chilonis (gi|700682216|), T. dendrolimi (gi|449043448|), T. ostriniae (gi|74483381|), T. brasiliensis (gi|74483391|), T. platneri (gi|1052465561|), T. pretiosum (gi|742527366|), T. evanescens (gi|742527376|), T. cacoeciae (gi|700676774|), and Scirpophaga incertulas (gi|633259618|). The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Maximum Composite Likelihood method [44] and are in the units of the number of base substitutions per site. Evolutionary analyses were conducted in MEGA7 [45].  Figure A4. The Trichogramma evolutionary history was inferred using the Neighbor-Joining method [42].