Introduction
Phytophagous insects often exhibit host specialization, where host-specialized populations assume greater fitness benefits on their natal hosts than on non-natal or novel hosts (Jaenike, Reference Jaenike1990). In fact, even a majority of generalists are known to exhibit host specialization to some degree (Futuyma & Moreno, Reference Futuyma and Moreno1988; Schoonhoven et al., Reference Schoonhoven, van Loon and Dicke2005; Sword et al., Reference Sword, Joern and Senior2005). Factors which may influence the degree of host specialization could be the genetic background of insect populations, local adaptation and experience (conditioning), differences in host availability in the landscape: spatially and temporally, nutritional quality of available host plants, and other ecological factors (Futuyma, Reference Futuyma1976; Bernays & Chapman, Reference Bernays and Chapman1994; McBride, Reference McBride2007). Host specialization, even in generalists, could lead to evolution of host biotypes (Carletto et al., Reference Carletto, Lombaert, Chavigny, Brevault, Lapchin and Vanlerberghe-Masutti2009).
Aphis gossypii Glover (Hemiptera: Aphididae) is a cosmopolitan and polyphagous pest. It utilizes plants belonging to Asteraceae, Cucurbitaceae, Malvaceae, and Solanaceae (Ebert & Cartwright, Reference Ebert and Cartwright1997; Blackman & Eastop, Reference Blackman and Eastop2000). A. gossypii exhibits considerable plasticity in host utilization, which allows it to use a variety of hosts in a landscape (Ebert & Cartwright, Reference Ebert and Cartwright1997; Wool & Hales, Reference Wool and Hales1997; Bancroft, Reference Bancroft2006). There are also numerous examples of host-specialized populations in the case of A. gossypii. For example, in Europe, A. gossypii from Chrysanthemum did not colonize cucumber and vice-versa (Guldemond et al., Reference Guldemond, Tigges and Devrijer1994). Similarly in China, A. gossypii from cucumber did not survive on cotton (Gao & Liu, Reference Gao and Liu2008). In addition to biological characterization of A. gossypii populations for their host specialization, several molecular studies have also revealed the existence of specialized host-adapted A. gossypii populations. Using microsatellite markers, Carletto et al. (Reference Carletto, Lombaert, Chavigny, Brevault, Lapchin and Vanlerberghe-Masutti2009) identified as many as five host biotypes (cucurbits, cotton, eggplant, potato, and chili) in A. gossypii populations collected from five different countries. The presence of multiple host biotypes, besides the commonly recognized cotton and cucurbit biotypes, was also documented by Gao & Liu (Reference Gao and Liu2008). Existence of various host biotypes in different geographical regions suggests that their evolution via host specialization may be contingent upon ecological characteristics, such as availability and composition of host and non-host plant species, host plant's phenological differences, climatic conditions, soil fertility, and agricultural practices (Funk & Bernays, Reference Funk and Bernays2001; Barman et al., Reference Barman, Parajulee, Sansone, Suh and Medina2012; Stotz et al., Reference Stotz, Suárez, Gonzáles and Gianoli2013).
Therefore, if host specialization is drastically influenced by the above-mentioned characteristics, then insect populations may not always follow the same pattern of host specialization throughout their distribution range (Fox & Morrow, Reference Fox and Morrow1981; Thompson, Reference Thompson2005). Different populations of the same herbivore species might exhibit specialization in some geographic locations while in other populations might be a generalist (Funk & Bernays, Reference Funk and Bernays2001; Gotthard et al., Reference Gotthard, Margraf and Rahier2004; Singer & McBride, Reference Singer and McBride2012). In several geographic regions around the world, A. gossypii populations were found to represent distinct host-specialized populations as revealed by several genetic and biological studies (Guldemond et al., Reference Guldemond, Tigges and Devrijer1994; Najar-Rodriguez et al., Reference Najar-Rodriguez, McGraw, Hull, Mensah and Walter2009; Agarwala & Choudhury, Reference Agarwala and Choudhury2013; Satar et al., Reference Satar, Kersting and Yokomi2013). In the agricultural landscape of the southeastern United States, especially in Georgia, both cotton and several cucurbitaceous vegetables are commonly grown in temporal and spatial proximities. While random migration of A. gossypii individuals between cotton and cucurbits is usually assumed, existence of host-associated biotypes in this species would undermine that assumption.
In this study, we assessed host specialization and host utilization plasticity of two A. gossypii populations. The two A. gossypii host-associated populations (HAPs) were field collected and later maintained on cotton and squash for several generations in the laboratory. First, we characterized the two HAPs, one on cotton and another on squash, using mitochondrial COI sequences. Later, fitness parameters of these two A. gossypii populations were evaluated on seven host plants; three host plants from Cucurbitaceae, two from Malvaceae, and two annual weed species belonging to two distant plant families (Amaranthaceae and Convolvulaceae). These plant species are typical representations of South Georgia farmscapes, where A. gossypii is highly prevalent and bound to encounter these selected plant species. Our objective was to measure the potential of the two HAPs in utilizing plant species other than their natal hosts. If HAPs cannot utilize alternate hosts, then there is a greater chance for that population to evolve into a distinct biotype. In contrast, presence of plasticity in host utilization in aphids would allow the population to feed opportunistically on an array of available hosts, and evolution of specialization will be constrained. Additionally, we evaluated phenotypical variations in two populations on their natal and alternate host species.
Materials and methods
Aphid populations
Approximately 100 A. gossypii adult apterae were collected from squash (Cucurbita moschata Duchesne ex Poiret) and cotton (Gossypium hirsutum L.) fields located at least 15 miles apart in Tifton, Georgia. These aphids were maintained separately on the originally collected host for multiple generations in greenhouse conditions (25–30°C and at 80–90% RH with a 14 h photoperiod). These two HAPs will be referred as either cotton-aphids or CAP and melon-aphids or MAP in this paper. The taxonomic identity of aphids, as A. gossypii, was confirmed by a hemipteran taxonomist, Dr Susan E. Halbert, at the Florida Department of Agriculture & Consumer Services in Gainesville, FL (voucher specimens: E2012-1467 and E2012-1470).
Host plants
A total of seven host species, two Malvaceaeous hosts (cotton, Gossypium hirsutum L. and okra, Abelmoschus esculentus L.), three Cucurbitaceaeous hosts (watermelon, Citrullus lanatus (Thunb.), squash, Cucurbita moschata Duchesne ex Poiret, and cucumber, Cucumis sativus L.), and two non-crop hosts (pigweed, Amaranthus palmeri S. Wats., Amaranthaceae; and morning glory, Ipomea alba L., Convolvulaceae) were used. All crop species (cotton, okra, watermelon, squash and cucumber) were grown from commercially available seeds, and experiments were conducted when plants were at 3–5 true leaf stage. Weed species (pigweed and morning glory) were grown from field collected seedlings devoid of any prior aphid infestations. These host plants were selected because they are prevalent in the farmscapes of southern Georgia.
Genetic diversity in aphid populations
In order to evaluate the genetic diversity of two host-associated (cotton and melon) populations of A. gossypii, four individuals from each population were randomly selected. Total DNA from individual aphids was extracted using Qiagen DNeasy kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. A universal primer pair C1-J-1718 (5′-GGAGGATTTGGAAATTGATTAGTTCC-3′) and 12-N-3014 (5′-TCCAATGCACTAATCTGCCATATTA-3′) was used to partially amplify the mitochondrial cytochrome oxidase subunit 1 (COI) gene following the polymerase chain reaction (PCR) methods and conditions specified in Simon et al. (Reference Simon, Frati, Beckenbach, Crespi, Liu and Flook1994). Electrophoresis was conducted on a 1% agarose gel stained with ethidium bromide (0.05 µg ml−1). The amplicons were purified using Qiaquick PCR purification kit (Qiagen, Valencia, CA). The purified PCR products were sequenced in both directions (Eurofins MWG Operon, Huntsville, AL). Sequences from a total of eight individual aphids (GenBank® accession number: KC610670- KC610677), four from each HAP, were selected for phylogenetic analysis. COI sequences of two voucher specimens of A. gossypii (GU591547 and JQ860257) were also included in the analysis. A. glycines partial COI gene sequence (EF467229) was included as the outgroup taxon. The sequences were aligned using Clustal-X and corrected by sight after setting the parameters for pairwise alignment (gap opening = 10; gap extension = 0.10), and multiple alignment (gap opening = 10; gap extension = 0.20; delayed divergence = 30%) (Hall, Reference Hall2001). Bayesian analysis was conducted in Mr. Bayes version 3.2 using the Monte Carlo Markovian Chain (MCMC) method (Ronquist and Huelsenbeck, Reference Ronquist and Huelsenbeck2003). The analysis used a general time reversible nucleotide substitution model with a gamma-distributed rate variation across sites along with a proportion of invariable sites. The phylogenetic tree was constructed using FigTree V1.4 (http://tree.bio.ed.ac.uk/software/figtree/).
Fitness parameters of A. gossypii
A pair of randomly selected apterous A. gossypii was placed using a camelhair paintbrush on the underside of a leaf of the respective host plants and confined with a clip cage. For each population, ten cages were placed on each host plant species. A total of 140 cages (2 HAPs × 7 host plants × 10 cages each) were placed. After 24 h of confinement, the females were removed, and only one freshly laid nymph was retained in each cage. The life history of this single aphid was monitored throughout the entire life cycle. For each HAP × host plant combination, only surviving nymphs from 140 cages were considered for statistical analysis (Table 1). To measure total fecundity, the nymphs laid by each subsequent generation adult were counted and removed daily. Fitness parameters such as nymphal period, adult period and total fecundity were measured and intrinsic rate of increase was estimated. Nymphal and adult periods in this context refer to the duration in days from the first day of larviposition to the first day of adulthood, and from the first day of adulthood to the completion of life cycle, respectively. Total fecundity represents the number of nymphs produced by an apterate adult throughout its adulthood. The intrinsic rate of increase (r m) for each aphid was calculated using the equation of Wyatt & White (Reference Wyatt and White1977):
$$r_m = 0.747\displaystyle{{{\rm lo}{\rm g}_eN_d} \over d}$$where N d is the number of nymphs produced during reproductive period, and ‘d’ is the pre-reproductive period in days. This experiment was repeated once with 10 cages for each host and each aphid population, thereby 280 (140 × 2) cages were used altogether.
Table 1. Host plant effects on developmental time and adult longevity of two A. gossypii HAPs.
*Cotton aphid – cotton associated population (CAP); melon aphid – melon associated population (MAP).
Phenotypical variation in A. gossypii
Approximately ten adults were placed on the undersurface of a leaf from seven different selected plant species. After 24 h, the adults were removed, and freshly laid nymphs were retained. These nymphs were allowed to feed on the respective host plants until they turned into adults, thus allowing completing an entire generation on the respective host plant. From each HAP × host combination, 30 A. gossypii adults were collected on the first day of their adulthood and preserved in 70% ethanol. A total of 420 A. gossypii adults (2 HAPs × 7 host plants × 30 adults each) were individually slide mounted after processing the samples following the methods described by Blackman & Eastop (Reference Blackman and Eastop2000). Each slide mount was photographed under the microscope (40 × magnification). Four different phenotypical variables: body length, head capsule width (measured distance between eyes), hind tibia length, and cornicle length were measured using Olympus DP2-BSW software (Olympus America, Inc., Center Valley, PA).
Statistical analyses
Data from two repeats of the fitness experiment were pooled for statistical analyses. Effects of aphid populations and host plants on A. gossypii nymphal and adult periods were analyzed using median one-way analysis using PROC NONPAR1WAY in SAS Enterprise version 6.1 (SAS Institute, Cary North Carolina). A. gossypii fitness parameters viz. fecundity and intrinsic rate of increase, and phenotypical parameters were evaluated using a generalized linear mixed model (PROC GLIMMIX) in SAS. Aphid populations and host plants were considered as fixed effects and replications as random effects. Treatment means were separated using Fisher's Least Squares Difference at a 95% confidence level.
Results
Genetic diversity in aphid populations
All the melon-aphids and cotton-aphids COI sequences in this study (Squash: KC610670-73; Cotton: KC610674-77) fell into a single A. gossypii clade. Mitochondrial COI sequences for the two HAPs in this study had maximum similarity with previously published A. gossypii COI sequences in the Genbank (GU591547, JQ860257, HQ528254, and FJ965680). This grouping was supported by 100% posterior probability value (Fig. 1). The MCMC tree also showed that all the A. gossypii were separated into one major clade from the out-group taxon, A. glycines (EF467229).
Fig. 1. Bayesian tree based on mitochondrial COI sequences of two host-associated A. gossypii populations and selected COI sequences of Aphis gossypii from GenBank.
Fitness parameters of A. gossypii
Nymphal period
The nymphal periods of both HAPs significantly varied with host plants (CAP: χ 2(1, 121) = 100.9, P < 0.0001; MAP: χ 2(1, 112) = 90.6, P < 0.0001; Table 1). Unlike other host plants, both populations did not have significantly different nymphal periods on cucumber and pigweed. Cotton-aphids reared on squash (7.5 days) and watermelon (7 days) took longer to become adults than when reared on cotton (5 days) and okra (4.5 days). A similar trend was observed with melon aphids. They took longer to become adults when reared on cotton (5 days) and okra (6 days) than on watermelon (4 days). Individuals from both populations developed rapidly into adults on cucumber (cotton aphid: 4 days; melon aphid: 4.5 days). However, the HAPs responded differently to one of the two weed species, morning glory. The nymphal period of cotton-aphids reared on morning glory (5 days) was shorter than the nymphal period of melon-aphids reared on the same host (7 days).
Adult period
Irrespective of rearing hosts, the adult periods for two aphid populations varied significantly (CAP: χ 2(1, 121) = 83.28, P < 0.0001; MAP: χ 2(1, 112) = 91.73, P < 0.0001). Adult periods of both populations also varied with host plants, except for watermelon. Cotton-aphids reared on cotton had the longest (21 days) and those reared on cucumber had the shortest (9 days) adult periods. Similarly, melon-aphids had the longest adult period on squash (19.5 days) and the shortest on cotton and okra (9 days each).
Total fecundity
Fecundity was affected by host association of populations (F (1, 223) = 96.41, P ≤ 0.0001). Irrespective of aphid population, host plants significantly affected aphid fecundity (F (6, 223) = 695.24, P ≤ 0.0001). Interactions between host plants and HAPs also significantly affected fecundity (F (6, 223) = 162.52, P ≤ 0.0001). Fecundity of both melon and cotton aphids was greater on squash and watermelon than on cotton and okra (Fig. 2). However, fecundity of melon-aphids was reduced on cotton and okra than on squash, melon, cucumber, pigweed, and morning glory.
Fig. 2. Mean ± SE values of fecundity in two host-associated A. gossypii populations as influenced by host plants. MAP- melon associated population, CAP- cotton associated population, Sqs- squash, Pgw- pigweed, Okr- okra, Mrg- morning glory, Wmel- watermelon, Cum- cucumber, and Cot- cotton.
Intrinsic rate of increase. Host association in a population by itself did not have any effect on intrinsic rate of increase (F (1, 231) = 3.52, P = 0.062). However, host plants influenced the intrinsic rate of increase for both aphid populations (F (6, 231) = 59.58, P ≤ 0.0001). Significant interactions between aphid populations and host–plant interactions were also observed (F (6, 231) = 138.55, P ≤ 0.0001). Intrinsic rate was reduced when individuals were exposed to their respective non-natal hosts. For example, when melon-aphids were reared on cotton or okra, their intrinsic rates of increase were less than when reared on hosts in the natal family, cucumber, squash, and melon (Fig. 3). Cotton aphids showed a similar trend in which their intrinsic rates of increase were higher on hosts in their natal family, cotton and okra, than on non-natal hosts such as squash and melon.
Fig. 3. Mean ± SE values of intrinsic rate of increase in two host-associated A. gossypii populations as influenced by host plants. MAP- melon associated population, CAP- cotton associated population, Sqs- squash, Pgw- pigweed, Okr- okra, Mrg- morning glory, Wmel- watermelon, Cum- cucumber, and Cot- cotton.
Phenotypical variation in A. gossypii populations
Regardless of host plants, melon-aphids had longer body length, hind tibia length, cornicle length and greater head capsule width than cotton-aphids (Fig. 4a–d). Furthermore, each of these phenotypical traits varied with host plant (Table 2). Significant interactions between HAPs and host plants were also observed for each phenotypical trait studied (Table 2). For instance, the melon aphids reared on squash and watermelon had greater body length, hind tibia length, cornicle length, and head capsule width than other host plants (Fig. 4a–d). The phenotypical characters of cotton aphids were in general smaller on cucumber than on most other hosts (Fig. 4a–d).
Fig. 4. Measurements (mean ± SE) of four phenotypical traits in A. gossypii populations as influenced by host plants. Parameters illustrated in figures are as follows: (a) body length, (b) head capsule width, (c) hind tibia length, and (d) cornicle length.
Table 2. Influence of host associated populations (HAPs) and host plants on phenotypical parameters of A. gossypii.
Discussion
We evaluated host utilization of two HAPs of A. gossypii using an array of aphid fitness and phenotypical traits. Our results demonstrate that these HAPs show plasticity in utilizing several hosts, thereby indicating absence of strict host specialization. Mitochondrial COI sequences from both of the aphid populations were identical, and did not reveal any evidence of genetic divergences between the two aphid populations. Mitochondrial genes often are used to detect recent divergence among taxa because of their faster rate of evolution than nuclear genes (Miyata et al., Reference Miyata, Hayashida, Kikuno, Hasegawa, Kobayashi and Koike1982; Simon et al., Reference Simon, Frati, Beckenbach, Crespi, Liu and Flook1994; Graur & Li, Reference Graur and Li1999; Wang et al., Reference Wang, Huang and Qiao2013; Song et al., Reference Song, Zhang, Li and Cai2016). Despite their increased rate of evolution, COI sequences often provide relatively low phylogenetic resolution at the population level, which consequently limits their use in validating lineages or biotypes. Further investigation of HAPs using higher resolution microsatellite markers, as shown by Brévault et al. (Reference Brévault, Carletto, Linderme and Vanlerberghe-Masutti2008), would have been more helpful to reveal genetic variability between cotton vs. squash-collected populations, if indeed there are differences. However, time and resource limitations precluded us from using higher resolution tools for this study. Apart from genetic divergence, other factors such as host range in the landscape and secondary endosymbiont composition could also influence aphids’ plasticity in host utilization (Fukatsu et al., Reference Fukatsu, Tsuchida, Nikoh and Koga2001; Medina et al., Reference Medina, Nachappa and Tamborindeguy2011; Brady & White, Reference Brady and White2013).
Two HAPs of A. gossypii showed improved overall development on their natal hosts and other closely related hosts within the same plant families. For example, the nymphal development time of both HAPs was shorter on their natal hosts and hosts within that same family than on hosts in the non-natal family. Also, the adult longevity was greater on their natal hosts and hosts within that family than on hosts in the non-natal family. Interestingly, the two HAPs performed intermediately on the two weed species (pigweed and morning glory), which suggests that in the absence of preferred crop hosts, aphids might be able to utilize alternate hosts. These results also indicate that these two HAPs are not host-specific biotypes and still retain their polyphagous abilities. In addition to the inter-population differences, there was substantial intra-population variation in both biological and phenotypic traits of the aphids utilizing different hosts within the natal host family. For example, melon-aphids utilizing squash lived significantly longer, produced more offspring, and were larger in size compared with melon-aphids utilizing watermelon and cucumber plants. Differential performance of A. gossypii on host plants within the same family can be attributed to variation in plant nutrient status and secondary metabolite profiles. Similar results were reported in the case of Uroleucon aphid while utilizing different Solidago hosts (Asteraceae), where variation in host nutritional quality influenced the performance of aphids (Moran, Reference Moran1981).
Although the two HAPs of A. gossypii in our study showed distinct patterns with respect to several biological and phenotypical traits following exposure to host plants in both natal and non-natal families, this does not necessarily warrant calling these two HAPs as host specialized populations or biotypes. This is especially true based on fecundity, which is a good measure of aphids’ performance on a particular host (Lamb et al., Reference Lamb, MacKay and Migui2009). In our study, we observed distinct trends for two HAPs. We found that the fecundity of melon- aphids was reduced by 89% on the non-natal host (cotton), while the fecundity of cotton-aphids rather increased by 66% on their non-natal host, squash. Thus, it appears that the melon-aphids are more specialized on cucurbit hosts (melon and squash) and less amenable to alternate host plants in non-natal families. On the contrary, the cotton-aphids did not appear to specialize on their natal hosts (Malvaceae); rather, they exhibited better reproductive performance on non-natal hosts such as watermelon and squash. An earlier study by Chen et al. (Reference Chen, Rahbe and Delobel1999) also reported similar results, in which melon aphids performed better on watermelon than cotton aphids. This enhanced performance was attributed to increased tolerance of melon aphids to pyrazole, an anti-herbivory secondary phloem metabolite. Thus, secondary metabolites, apart from nutrients such as amino acids, may have contributed to the differential responses of aphids to natal and non-natal hosts thereby affecting aphid growth and fitness.
Phenotypical measurements such as body length, tibial length, and cornicle length have been used to describe variation in aphid populations utilizing different host plants (Margaritopoulos et al., Reference Margaritopoulos, Tsitsipis, Zintzaras and Blackman2000; Lee & Lee, Reference Lee and Lee2013). Such phenotypical variations in populations experiencing different feeding environments may directly affect aphid fitness (Honek, Reference Honek1993; Via & Shaw, Reference Via and Shaw1996; de Kogel et al., Reference de Kogel, Bosco, van der Hoek and Mollema1999; Pereira & de Paula, Reference Pereira and de Paula2009). Melon-aphids were significantly larger than cotton-aphids regardless of developmental host plants. The inherent size difference in two HAPs might be due to the quality of original feeding hosts. This may have led to the larger body size in melon-aphids along with their higher reproductive performance compared with aphids that developed on cotton. Similar observations were made by Wool & Hales (Reference Wool and Hales1997) in Australian A. gossypii populations, which were larger in size upon rearing on cucumber and rockmelon compared with aphid populations reared on cotton, broad bean, and eggplant. Relationship of host plant quality and aphid's reproductive performance was also validated in pea aphids, Acyrthosiphon pisum (Harris), which produced more offspring on Vicia faba L., a supposedly superior host than Lathyrus pratensis L. (McLean et al., Reference McLean, Ferrari and Godfray2009). Interestingly, this host effect on A. gossypii is observable within a generation. In this study, when larger sized melon-aphids were exposed to either cotton or okra plants as developmental hosts, a reduction in size was observed. Similar trends were observed in the case of two other phenotypical traits evaluated, hind tibia and cornicle length.
Plant quality and secondary metabolite profiles play a significant role in aphid's performance and phenotypic development (Moran, Reference Moran1981; Douglas, Reference Douglas1993; Awmack & Leather, Reference Awmack and Leather2002; Powell et al., Reference Powell, Tosh and Hardie2006). In addition to these plant attributes, the maternal effect –progeny's performance as a result of the mother's experience is a critical determinant of aphid fitness (Mousseau & Dingle, Reference Mousseau and Dingle1991; Fox et al., Reference Fox, Waddell and Mousseau1995). A strong maternal effect on the subsequent offspring generations could be expected as aphids have overlapping telescopic generations (Dixon, Reference Dixon1998). Several studies have examined as to whether the maternal effect on subsequent offspring performance plays a significant role (Via, Reference Via1991; Zehnder & Hunter, Reference Zehnder and Hunter2007; McLean et al., Reference McLean, Ferrari and Godfray2009; Tariq et al., Reference Tariq, Wright and Staley2010). Aphids may need to be followed for several generations in the novel environment in order to eliminate any bias associated with the presence of maternal effects (Mousseau & Dingle, Reference Mousseau and Dingle1991; Olivares-Donoso et al., Reference Olivares-Donoso, Troncoso, Tapia, Aguilera-Olivares and Niemeyer2007). To separate maternal effects from host effects in A. gossypii populations, future studies could investigate how long (number of generations) the mother has to be on a specific non-natal host in order to observe the maternal effect on its offspring.
Our study further provides insights into the potential role of annual weed species as alternate hosts on the biology of A. gossypii in the context of agricultural landscape. Among the several dominant weed species in farmscapes of the southeastern United States, Pigweed, and morning glory are prominent (Webster & MacDonald, Reference Webster and MacDonald2001). In a given year, the availability of cultivated crop may vary temporally and spatially within these farmscapes. Also, several other factors such as deteriorating quality of senescing crops, overcrowding of aphid populations, and pesticide applications may drive individuals to seek alternate hosts, which could be one or more annual weed species. This study clearly shows that regardless of natal hosts, A. gossypii can sub-optimally exploit these two annual weed species, pigweed, and morning glory. While there are several reports of A. gossypii using perennial plants such as Hibiscus for overwintering purposes (Liu et al., Reference Liu, Zhai and Zhang2008a, Reference Liu, Zhai, Zhang and Gub; Razmjou et al., Reference Razmjou, Vorburger, Moharramipour, Mirhoseini and Fathipour2010), records are limited when it comes to annual non-crops serving as alternate hosts (Perng, Reference Perng2002; Margaritopoulos et al., Reference Margaritopoulos, Tzortzi, Zarpas and Tsitsipis2009). In Taiwan, Perng (Reference Perng2002) documented that among the four weed species, Solanum nigrum L. (black nightshade) was a better host of A. gossypii than Ageratum houstonium Mill., Bidens pilosa L., and Spermacoce latifolia (Aubl.). Not all these species may be equally important for every geographical region. Nevertheless, regionally dominant weed hosts could provide important information and aid in understanding the dynamics of aphid populations in an agricultural landscape.
Concluding remarks
We observed that the two HAPs of A. gossypii in this study exhibited plasticity in host utilization. The HAPs were not necessarily specialized or genetically diverged. The lack of severe fitness costs on non-natal hosts suggests that divergence into host-specific biotypes is unlikely. Despite this, identifying the common patterns associated with host plant utilization by multiple HAPs in a farmscape/landscape setting could potentially be helpful in understanding dispersal patterns and in designing the pest management strategies. If such HAPs evolve in to specific biotypes, then pest management strategies could be tailored for each biotype. An example being the case of whitefly biotypes B and Q, wherein the Q biotype has increased capabilities to develop resistance to insecticides than B, and typically B is more competitive than Q (Horowitz & Ishaaya, Reference Horowitz and Ishaaya2014). Timely knowledge on host specialized populations/lineages/biotypes and understanding their dispersal patterns would serve as valuable information from a commercial agriculture standpoint.
Acknowledgements
The authors wish to express their gratitude to Ms. Julia Kitchens for technical help rendered during the course of the experiment. The authors also acknowledge Mr. Simmy McKeown and Ms. Sheran Thompson for technical help.
