Effect of nutritious and toxic prey on food preference of a predaceous ladybird , Coccinella septempunctata ( Coleoptera : Coccinellidae )

We investigated the predatory potential and food preference of different life stages of Coccinella septempunctata L. for a nutritious aphid (mustard aphid, Lipaphis erysimi) and toxic aphid (cabbage aphid, Brevicoryne brassicae). We provided all the life stages of C. septempunctata with either L. erysimi or B. brassicae and found that the second, third and fourth instar larvae and adult females of this predator consumed daily greater numbers of L. erysimi. However, the fi rst instar larvae and adult males consumed similar numbers of both of these aphids. In choice condition, each larva, adult males and females were each provided separately with a mixed aphid diet in three proportions (i.e. low: high, equal: equal and high: low densities of L. erysimi: B. brassicae). We hypothesized that life stages of C. septempunctata will prefer L. erysimi regardless of its proportions. Laboratory experiments supported this hypothesis only at the adult level in terms of high values of β and C preference indices. However, it rejects this hypothesis at the larval level, as larvae preferred B. brassicae when provided with certain combinations and showed no preference in a few combinations. We infer that mixtures of nutritious and toxic aphids may enable this ladybird to overcome any probable nutritional defi ciency and/or reduce the toxicity of a toxic diet, especially for the larvae. Results of the treatment in which a high proportion of B. brassicae were consumed along with fewer L. erysimi indicates that a mixed diet could be better for the development of immature stages of C. septempunctata. * Corresponding author; e-mail: ahmadpervez@yahoo.com INTRODUCTION Predatory potential and food preference are important aspects of a predator’s biocontrol potential. Predators prefer certain prey (Sundell et al., 2003), however, the many species of prey in nature affects their ability to attack their preferred prey (Dicke et al., 1989). Testing their food preference by presenting two species of prey in equal numbers separately or together is not enough, as mixtures of different proportions of species of prey could infl uence a predator’s choice of food (Murdoch, 1969; Chesson, 1984). Predators show different behavioural responses, such as (i) preference for a particular species (ii) no preference or null switching, (iii) preference for the most abundant prey, i.e. switching behaviour, and (iv) preference for the less abundant prey, i.e. anti-switching behaviour. Seemingly, prey preference is dependent on multiple factors, viz. prey availability, prey specifi city, prey capture time, prey handling time, nutritive value, palatability, promoting fi tness, lesser costs of predation, etc. (Hodek & Evans, 2012). Eur. J. Entomol. 114: 400–406, 2017 doi: 10.14411/eje.2017.051

Most previous studies on prey preferences of predaceous ladybirds (Coleoptera: Coccinellidae) use single species of prey separately and record their effect on consumption and fi tness (Omkar & Bind, 1998;Omkar et al., 1999;Omkar & Pervez, 2001;Pervez & Omkar, 2004;Omkar & Mishra, 2005).However, a preference test is more appropriate when there are several sources of food in the same microhabitat, i.e. a cafeteria setup (Ferrer et al., 2008;Nedved & Salvucci, 2008;Šenkeříková & Nedvěd, 2013).In this situation, both adult males and females of the Harlequin ladybird, Harmonia axyridis (Pallas) exhibit different types of response to three different relative abundances of prey, as males prefer certain prey and females do not (Soares et al., 2004).Hodek & Evans (2012) elucidated the concepts of ladybirds' accepted food in terms of "essential food" (supports both oviposition and development) and "alternative food" (only ensures survival).Essential food provides necessary nutrients needed for egg production and oviposition (Evans & Gunther, 2005), and hence we consider it as nutritious.Alternative food perhaps lacks these essential Effect of nutritious and toxic prey on food preference of a predaceous ladybird, Coccinella septempunctata (Coleoptera: Coccinellidae) INTRODUCTION Predatory potential and food preference are important aspects of a predator's biocontrol potential.Predators prefer certain prey (Sundell et al., 2003), however, the many species of prey in nature affects their ability to attack their preferred prey (Dicke et al., 1989).Testing their food preference by presenting two species of prey in equal numbers separately or together is not enough, as mixtures of different proportions of species of prey could infl uence a predator's choice of food (Murdoch, 1969;Chesson, 1984).Predators show different behavioural responses, such as (i) preference for a particular species (ii) no preference or null switching, (iii) preference for the most abundant prey, i.e. switching behaviour, and (iv) preference for the less abundant prey, i.e. anti-switching behaviour.Seemingly, prey preference is dependent on multiple factors, viz.prey availability, prey specifi city, prey capture time, prey handling time, nutritive value, palatability, promoting fi tness, lesser costs of predation, etc. (Hodek & Evans, 2012).
(size as above).At egg hatch the neonates were transferred to prey infested potted plants in acrylic cages (45 × 45 × 45 cm).For the stock maintenance of the two species of aphids: L. erysimi and B. brassicae, two plants: mustard (Brassica campestris) and cabbage (Brassica oleracea) were grown in pots in the laboratory, each replicated many times.We simultaneously maintained parthenogenetic cultures of these aphids on mustard and cabbage plants in the laboratory under the above abiotic conditions.

Predatory potential of C. septempunctata
We determined the predatory potential of the larvae and adults of C. septempunctata fed either L. erysimi (Le) or B. brassicae (Bb) in order to quantify aphid consumption in a no choice condition (both daily and total prey consumption).For this purpose, we kept newly hatched fi rst instar larva of C. septempunctata in a glass beaker (11.0 cm in height and 9.0 cm in diameter) containing a piece of the host plant infested with 50 third instar nymphs of L. erysimi (host plant as cited above).The beaker was covered with muslin cloth fastened with a rubber band and kept in an Environmental Test Chamber (conditions as above).We recorded moulting every 8 h.After 24 h, the beaker was removed from the chamber and the number of live aphids counted in order to quantify the number of aphids consumed.If moulting had not occurred, the aphids were replenished (same number of aphids as provided the previous day) and this was repeated daily until moulting occurred.At this moult prey consumption was recorded (as above).Thereafter, we provided 100, 200 and 200 third instar nymphs of L. erysimi daily to second, third and fourth instar larvae, respectively, until they moulted / pupated and the number of prey consumed by each of these instars was recorded.Similarly, we provided 300 aphids daily to newly emerged adult males and females until they died and recorded the number of prey consumed daily (n = 10).These experiments were repeated using B. brassicae as prey.
We tested the data on prey consumption for normality using the Kolmogorov-Smirnoff test and homogeneity of variance using Bartlett's test in statistical software, SAS Ver.9.0.The data on prey consumption by different stages of C. septempunctata were subjected to the two sample t-tests of means in SAS 9.0.The data were also subjected to Two-way ANOVA in SAS 9 with "species" (two levels) and "stage" (six levels) as independent variables and the daily and total prey consumption as dependent variables.

Food preference of C. septempunctata
We offered aphids, Le and Bb in three ratios, to fi rst (i.e.Le:Bb 12:38,25:25 and 38:12), second (i.e.Le:Bb 25:75, 50:50, 75:25), third (i.e.Le:Bb 50:150, 100:100, 150:50) and fourth (i.e.Le:Bb 50:150, 100:100, 150:50) instar larvae, adult males (i.e.Le:Bb 75:225, 150:150 and 225:75) and females (i.e.Le:Bb 75:225, 150:150 and 225:75) of C. septempunctata and recorded the numbers of each species of aphid consumed after 24 h, using each life stage only once.We calculated Manly's preference index (Manly, 1972) for each treatment using the formula, β = log (N A / r A ) / [log (N A / r A ) + log (N B / r B )], in which N A and N B are the numbers of prey A and prey B offered to a predator and r A and r B are the numbers of unconsumed prey.This index overcomes the problem of an error resulting from prey depletion, i.e. it is applicable in those experiments in which killed prey are not replaced (Cook, 1978;Sherratt & Harvey, 1993).If β is close to 1, the predator prefers prey A and if close to 0, prey B is preferred.An index value close to 0.5 indicates no preference.We tested β for signifi cant difference from a value of (0.5) using a one sample t-test of the results of each treatment in statistical software, MINITAB 13.0.Prey preference was also analyzed using the C index, i.e.C = (E A × N B ) / (E B × N A ) (Sherratt & Harvey, 1993), where E A and E B are nutrients and/or contains allelochemicals that make it toxic to ladybirds (Hodek & Evans, 2012).Generalist predators consume alternative prey in order to avoid starving when essential food is scarce (Sherratt et al., 2004).However, certain predators may readily attack alternative prey even when nutritious prey are abundant, as the former may be easier to capture (Lang & Gosdl, 2001;Provost et al., 2006).
It is likely that when providing essential and alternative foods simultaneously, ladybirds will prefer to consume the former, as it is positively associated with development and reproduction (Omkar & Mishra, 2005).Mixing certain non-insect foods and/or less abundant prey with the preferred abundant prey may enhance the fi tness of predaceous insects (Moser et al., 2008;Lundgren et al., 2009).Hence, it would be interesting to know, how ladybirds respond when provided with both toxic and nutritious prey simultaneously.Keeping in view, the earlier concept of food preference and food availability, it is likely that life stages of ladybirds are more likely to prefer nutritious over toxic prey.
Coccinella septempunctata L. is an aphidophagous ladybird with a wide prey range (Evans, 2000;Omkar & Pervez, 2002, 2004;Hodek & Michaud, 2008).However, it is seasonally synchronized with the mustard aphid, Lipaphis erysimi (Kalt.) in North India, which is one of its preferred aphids (Omkar et al., 1999;Ali & Rizvi, 2007).Omkar & Srivastava (2003) found L. erysimi to be the most nutritious of six species of aphids tested, in terms of the growth, development and reproduction of C. septempunctata.In contrast, this ladybird avoids eating the cabbage aphid, Brevicoryne brassicae L. (Bilashini et al., 2007;Khan & Khan, 2002;Papachristos et al., 2015), as it sequesters glucosinolates from its host plant and uses them as a defence against predators (Pratt, 2008;Kos et al., 2011Kos et al., , 2012)).B. brassicae is even labelled as a rejected prey because of its unpalatability due to its waxy surface (Hodek & Evans, 2012).Hence, it is likely that C. septempunctata will prefer L. erysimi over B. brassicae regardless of whether the former prey is less, equal or more abundant in a mixture of prey.We hypothesized that C. septempunctata will prefer L. erysimi in all proportions of aphid combinations.We also aim to study the effect of the presence of nutritious prey on the toxicity of toxic prey for this ladybird.Hence, the present investigation determines the effect of nutritious and toxic prey on the predatory potential and food preference of larvae and adults of C. septempunctata.

Stock maintenance
We collected adults of C. septempunctata from agricultural fi elds near the university campus, Jammu, India (33.7782°N, 76.5762°E) and brought them to the laboratory.Ten pairs of adults were each kept separately in Petri dishes (2.0 cm height × 9.0 cm diameter) and allowed to mate.Thereafter, female ladybirds were each isolated in separate Petri dishes and their egg laying recorded under controlled conditions (25 ± 2°C, 60% RH, and 16L: 8D photoperiod) in an Environmental Test Chamber (Remi, India) and the eggs were transferred to other Petri dishes the number of prey A and prey B consumed.C value between 0 and 1 indicates a preference for prey B and a value of more than 1 indicates a preference for prey A. We determined whether the C-index recorded in each treatment was signifi cantly different from one using a one sample t-test in MINITAB 13.0.The Cvalue analysis confi rmed the food preference of the predator.We subjected the data on number of prey consumed to Wilcoxon's matched-pairs signed rank test, and the proportion of each prey consumed to two sample t-tests in SAS Version 9.0.
Both the male and female adults consumed a signifi cantly greater proportion of L. erysimi than B. brassicae in all the mixtures (Fig. 2).This resulted in signifi cantly higher β and C indices and greater positive t-values confi rming a preference for L. erysimi (Table 1).The fourth instar larvae signifi cantly preferred B. brassicae when provided with the 50:150 (Le:Bb) mixture, and L. erysimi when provided with the other two mixtures.The third instar larvae preferred L. erysimi only when provided with the 150:50 (Le:Bb) mixture and showed no preference when provided with the other two mixtures.Second instar larvae preferred B. brassicae when provided with the 75:25 (Le:Bb) mixture, and both their β and C indices had signifi cant negative t-values (Table 1).First instar larvae signifi cantly preferred L. erysimi when provided with the 12:38 (Le:Bb) mixture and B. brassicae when provided with the 38:12 (Le:Bb) mixture (Fig. 1; Table 1).Table 1.Mean values of β and C recorded for C. septempunctata provided with different mixtures of the aphids L. erysimi and B. brassicae.This predator prefers L. erysimi if β is close to 1 and B. brassicae if β is close to 0, and exhibits no preference if β is close to 0.5.C index more than 1 indicates a preference for L. erysimi and between 0 and 1 a preference for B. brassicae.

DISCUSSION
In no choice experiments most life stages of C. septempunctata consumed a greater number of L. erysimi than B. brassicae.This is largely ascribed to the high protein content of L. erysimi along with its seasonal synchrony with C. septempunctata (Atwal & Sethi, 1963;Omkar & Srivastava, 2003).In north India, three species of aphids: L. erysimi, Myzus persicae (Sulzer) and B. brassicae, infest Brassica crops in succession from December to March (Bilashini et al., 2007).Infestations of L. erysimi in the fi elds and the arrival there of C. septempunctata are synchronized (Sharma et al., 1997).Dense infestations of L. erysimi give rise to an abundance of C. septempunctata, and the decline of L. erysimi in February is associated with the appearance of the fi rst colonies of B. brassicae, which is a poorer quality prey (Bilashini et al., 2007;Bilashani & Singh, 2009).Coccinella septempunctata is strongly correlated with L. erysimi in terms of its abundance and weakly so with B. brassicae (Bilashini et al., 2007).Reduced consumption of B. brassicae recorded in our study could explain the low numbers of C. septempunctata in this crop at this time (Ahuja et al., 2010).
The signifi cant main effects of "species" on daily and total prey consumption recorded in the no choice experiment indicate that larvae and adults C. septempunctata responded differently to the two aphids provided and prefer L. erysimi.Similarly, the signifi cant main effect of "stage" on prey consumption reveals that different larval stages and adults have different quantitative dietary requirements.Prey consumption progressively increased with age and is associated with the increased food requirements for sustaining their growth and metabolism as they increased in size.In addition, the trend with increase in size was independent of the species of aphid provided.However, owing to the toxic nature of B. brassicae, it was expected that fewer of this species would be consumed.The fi rst instar probably needs a certain biomass of prey in order to complete its development and as its sensory ability is probably poorly developed compared to that of following instars it consumes the same numbers of both species of aphid.Female ladybirds are more voracious than male ladybirds, as more prey are needed to maintain their bigger body size and develop their gonads (Rhamhalinghan, 1987;Lucas et al., 1997).
The lower consumption of B. brassicae by most of the stages of C. septempunctata is largely ascribed to glucosinolates, which the aphid sequesters from its host plant and uses as a defense against predaceous ladybirds (Pratt, 2008;Kos et al., 2011Kos et al., , 2012)).In addition, endogenous myrosinase in B. brassicae cumulatively forms hydrolytic products with glucosinolates that could be even more toxic for aphid predators (Francis et al., 2002;Kos et al., 2011Kos et al., , 2012)).Furthermore, powdery wax on the surface of the body of B. brassicae possibly makes it less attractive and palatable for C. septempunctata.
The choice experiment revealed that overall the larvae, and adult males and females of C. septempunctata prefer L. erysimi.This partially supports our hypothesis that C. septempunctata will prefer L. erysimi when provided with different mixtures of the two aphids.Mostly, the larvae show no preference when provided with equal numbers of each species of aphid and in a few instances, they consume a greater proportion of B. brassicae than L. erysimi.Nedved & Salvucci (2008) found that C. septempunctata consumes the toxic aphid, Aphis sambuci L. at twice the rate of two essential aphid species: Acyrthosiphon pisum (Harris) and Aphis philadelphi.Similarly, adult male, H. axyridis consume more of the toxic aphid, A. sambuci than of the suitable prey Aphis fabae cirsiiacanthoidis (syn. A. philadelphi) (Šenkeříková & Nedvěd, 2013).First instar larvae of Adalia bipunctata (L.) attack and consume more of the toxic aphid, Aphis craccivora Koch than the nutritious aphid, A. pisum, whilst fourth instar larvae exhibited a null switching response by choosing none (Ferrer et al., 2008).We also recorded for larvae, that the toxic and rejected prey, B. brassicae (Hodek & Evans, 2012) was readily consumed and even preferred on a few occasions.Hence, providing mixtures of nutritious and toxic aphids might overcome any nutritive defi ciency and/or decrease the toxicity of a toxic diet for larvae.Our results also support the fi nding of Soares et al. (2004), in which adult females of H. axyridis showed a null switching response to a mixed aphid diet.The inferred dietary self-selection behaviour of H. axyridis probably accounts for a mixed prey diet being more suitable than a single prey diet for this predator (Soares et al., 2004).Thus, it is likely that when L. erysimi is scarce the need for protein in the diet (Atwal & Sethi, 1963) is met by consuming B. brassicae, especially when it is abundant.Rana et al. (2002) comment that ladybirds may prefer toxic prey over essential prey if continuously reared on the former for a few generations.Such fi ndings create doubts about the preference of ladybirds for the most nutritious prey and threaten the validity of the established defi nitions of preferred, essential, alternative and toxic foods (Hodek & Honek, 1996;Hodek & Evans, 2012).We conclude that B. brassicae is not toxic for or rejected by C. septempunctata and a mixed diet of B. brassicae and L. erysimi could be benefi cial for the development of this ladybird.In addition the different life stages of C. septempunctata (i) quantitatively consume more L. erysimi than B. brassicae when provided with each of these aphids separately, and (ii) adults prefer L. erysimi regardless of the proportion of this species in a mixed diet, (iii) larvae did not show a prey preference on most occasions and readily attacked both of the aphids provided.

Fig. 1 .
Fig. 1.Daily consumption of two species of aphids by the different stages of C. septempunctata in no choice experiments.Error bars denote standard error.Different letters indicate that the data is signifi cant.

Fig. 2 .
Fig. 2. Proportions of two species of aphids, L. erysimi (grey columns) and B. brassicae (black columns), consumed in choice experiments by the different stages of C. septempunctata.The bars at the tops of the columns are standard errors.