Action threshold development in cabbage pest management using 1 synthetic and botanical insecticides

25 As synthetic insecticides can have environmentally detrimental side effects, it is desirable to 27 limit their use while still achieving good marketable yield. One approach is to apply 28 pesticides only when needed, as determined by an action threshold (AT), defined as the 29 number of pests per crop plant or damage intensity at which application is recommended. 30 Another approach is to adopt alternative pesticides, which can also be applied according to 31 ATs. Here, ATs are developed in cabbage pest management using both approaches against the 32 moths Plutella xylostella (L.), Helicoverpa armigera (Hübner) and Spodoptera litura (F.) and 33 the aphid Brevicoryne brassicae (L.). Action thresholds were derived using fixed spraying 34 regimes for the synthetic insecticides (imidacloprid and Voliam Flexi) and for azadirachtin, a 35 neem-derived botanical. Botanical insecticide was as effective as synthetics in suppressing 36 pests and protecting yield. For synthetics, derived ATs are 40 individuals per plant for B. 37 brassicae , 0.3 larvae for P. xylostella and 0.2 medium-sized larvae for H. armigera and for S. 38 litura . For H. armigera and S. litura , negative relationships between marketable yield and 39 pest were found when larvae were medium or large sized, but not when larvae were small. 40 Compared to synthetics, benefits of using neem formulations include higher action thresholds 41 against P. xylostella (0.6/plant) and H. armigera (0.4/plant) and an oviposition deterrent effect 42 against S. litura . Although regional limits may apply to the accuracy of any ATs derived, the 43 approach used towards their establishment is simple and transferable to other agricultural 44 regions and crops. 45

). When experimental cabbages were infested with lepidopterans 173 only, Voliam Flexi was sprayed every 5 th , 10 th or 15 th day; and when only aphids and no 174 lepidopterans were present, plots that were due to be sprayed with Voliam Flexi were instead 175 sprayed with imidacloprid every 7 th , 14 th or 21 st day. However, when lepidopterans and aphids 176 infested cabbages simultaneously, three plots sprayed with Voliam Flexi were also sprayed 177 with imidacloprid. The shortest interval by either insecticide represents the normal practice of 178 the local growers. Further experimental treatments were: spraying botanical NeemAzal or 179 NSE at weekly intervals (both were collectively used for aphids and/or lepidopteran pests and 180 the inclusion of control plots. This was the core protocol employed during both study years.

181
For both neem-derived botanicals, fortnightly spray regimes in the first year were also trialed 182 but, as these transpired to be less effective than their weekly-applied counterparts (see 183 Results), they were not included in the second year of trials.

184
Sampling and yield assessment 185 Pest sampling was carried out on a weekly basis between initial pest appearance and the time 186 of crop harvest. Incidence of attack by lepidopteran larvae and aphids was measured as opened to obtain information on pest species present whereas P. xylostella and B. brassicae 200 were easily observed within head leaves. For deriving action thresholds, the acceptably 201 marektability criterion was set at 90%, following local grower practice for commercial sale.

224
Percent marketable yield obtained from insecticide treatments were arcsine-square root 225 transformed prior to ANOVA. Relationships between pests (total numbers or larval size class) 226 and marketable yield were assessed using regression analysis. All data analyses were 227 performed using the SPSS software package (version 21).

229
The guild of pest insects associated with cabbage comprised two aphid and six lepidopteran 230 species (Fig. 1a,b). In terms of abundance and persistence, the dominant lepidopterans were S.   (Fig. 1a,b). Spodoptera litura and H. armigera constituted the most persistent pest 233 complex, present mostly concurrently in October and November and again from February to 234 early-June (Fig. 1a). The pest complex was diversified by the appearance of P. xylostella and 235 B. brassicae from early-January and infestation continued until April or May (Fig. 1a,b). 236 Remaining pests (T. orichalcea, P. brassicae, S. exigua and M. persicae) were infrequently 237 present and at lower densities and were thus considered as minor pests (Fig. 1a,b). As minor 238 pests can be managed as an indirect consequence of the management employed for major 239 pests, major pests were the focus for presenting results and deriving ATs.  241 The overall composition of pest (species and numbers) present was influenced by the date of October and January whereas S. litura was abundant when cabbages were grown either in 253 September or from February to March (Fig. 1c). 255 The effect of insecticides, sampling dates and their interactions were typically significant for 256 B. brassicae, S. litura, P. xylostella and H. armigera, and also for minor pests (Table 2).

257
Brevicoryne brassicae was present in eight out of 15 trials, whereas S. litura was present in 258 six of these. However, H. armigera and P. xylostella, were found in almost all trials (14/15; 259 Table 2). Patterns of weekly abundance of each of these four pests are illustrated in     35% and 98%, respectively (Fig. 2). In a few trials, S. exigua larvae were recovered from 305 infested heads but their proportional loss was not more than 10% in untreated plots (Fig. 2).

306
Although spraying synthetics and neem-derived compounds protected yield losses, the 307 acceptable criterion for marketability (set at 90% of yield) was only achieved when plots were 308 sprayed with Voliam Flexi every 5 th day or when sprayed weekly with NeemAzal (Fig. 2).

309
Spraying Voliam Flexi every 10 th day only sometimes produced yields that were at least 90% 310 marketable (Fig. 2). 312 The identification of predictors of yield is fundamental to the derivation of ATs. The 313 companion study on cauliflower crops found that peak pest infestation, across an extended  species within the complex of pest herbivores (Saeed et al. 2015) and, in consequence, the 365 optimal insecticide application program. In the case of vegetable production, market prices 366 can vary on a monthly basis, depending upon consumer demand (generally higher for the first 367 and last crops of a season). Thus, planting date adjustment is unlikely to be adopted by 368 commercial scale growers but may be used by subsistence growers and also serves to identify 369 periods when pests are likely to become abundant and thus the frequency of control required.

399
If was found that more frequent spraying with synthetics always resulted in high yield (>90%) 400 and that less frequent spraying did not, although at most times pest densities were similar (as   Table 2 gives planting dates at each site).