Diversity, Host Plants and Potential Distribution of Edible Saturniid Caterpillars in Kenya

Simple Summary Edible insects are a traditional food source with economic benefits in sub-Saharan Africa. Caterpillars are the most popular edible insects in this region. We focus on caterpillars in the family Saturniidae. Saturniids are big colorful caterpillars with spines on their bodies, usually found in shrubs and trees. They are rich in proteins, vitamins, and minerals. Despite their economic importance, little is known about their diversity, host plants, distribution, and potential effect of climate change on edible saturniid caterpillars in Africa. The aim of this study is to identify edible saturniids, their host plants, their current distribution and to predict the possible effects of climate change on their distribution. We documented seven species of edible saturniids namely Gonimbrasia zambesina, Gonimbrasia krucki, Bunaea alcinoe, Gonimbrasia cocaulti, Gonimbrasia belina, Gynanisa nigra and Cirina forda. These caterpillars mostly occur twice a year during the rainy seasons and feed on specific host plants. Predictive distribution models revealed that B. alcinoe, and C. forda are mostly found in tropical and sub-tropical regions in Africa. However, climate change could cause a slight decrease in their population by the year 2050. This information will guide conservation efforts and ensure sustainable use of edible saturniid caterpillars as food. Abstract The promotion of edible insects, including saturniid caterpillars as potential food source is widely gaining momentum. They are adequately rich in nutrients such as proteins, amino acids, fatty acids, and micronutrients. Despite saturniids being a traditional food source with economic benefits, information on their diversity, host plants and their potential distribution in Africa are lacking, which this study seeks to address. Edible saturniids and their host plants were characterized using specific primers (LepF1/LepR1 and 3F_KIM_F/1R_KIM_R, respectively). Maximum entropy (MaxENT) and GARP (genetic algorithm for ruleset production) models were used to characterize the potential distribution of commonly consumed saturniids under current and future climate scenarios. Seven species of saturniids were recorded from 11 host plants in Kenya: Gonimbrasia zambesina, Gonimbrasia krucki, Bunaea alcinoe, Gonimbrasia cocaulti, Gonimbrasia belina, Gynanisa nigra and Cirina forda. Two morphotypes of G. zambesina and B. alcinoe were recorded. These saturniid caterpillars occur twice a year except for G. cocaulti. Predictive models revealed that tropical and subtropical regions were potentially suitable for B. alcinoe and C. forda. The information generated from this study would be important to guide conservation efforts and their sustainable utilization as food in Africa.

Field-collected Saturniid larvae were reared at 12 h:12 h photoperiod at 25 • C in Perspex cages (50 × 50 × 50 cm) with nets on the sides for ventilation. Cages were protected from ants and crawling insects with traps containing water placed below its metallic stand. The saturniid larvae were fed on twigs of their respective host plant from the field. The twigs were placed in a container as bouquet with stem immersed in water fastened by wet cotton wool. The twigs were changed daily to keep them fresh. The larvae fed until they reached the pre-pupal stage, when they stopped feeding, reduced movement, and moved to the floor of the cage ready to burrow and pupate. The pre-pupae were placed on moist sterile sawdust in plastic trays and allowed to burrow into the sawdust to pupate as they burrow in soil in the wild. The saw dust was kept moist by sprinkling water daily. The trays with the pupae were placed in Perspex cages (50 cm × 50 cm × 50 cm) awaiting emergence of adult moths.

Morphological Identification
The adult moths were killed by placing them in a container and freezing them at −5 • C. They were stretched out, pinned, allowed to dry, and labeled before morphological identification. Identification was carried out using published keys [12] and crosschecked with reference voucher specimens at National Museums of Kenya (NMK) collection and pictures from available literature [40,41] by Mr. Alex Musyoki, Mr Ashikoye Okoko and Dr. Esther Kioko, NMK. Voucher specimens are deposited at the Biosystematics Unit, icipe. Host plants were identified at the Kenya Forest Research Institute (KEFRI) by an experienced plant taxonomist using available literature [42].

Molecular Identification 2.4.1. Tissue Preparation, DNA Extraction and Quantification
Leaf samples of each host plant of saturniid caterpillar were carefully washed with tap water, rinsed with distilled water and dried with paper towel. They were then cut into 0.2 g of leaf sample small pieces with a sterile blade and placed in a 2 mL tube containing ceramic beads, lysis buffer PA1 (Bioline, London, UK) and RNAse A and crushed for 3 min in a Tissue lyser II (Qiagen, Germantown, MD, USA). Plant genomic DNA was extracted using Isolate II Plant DNA extraction Kit (Bioline, London, UK) as per the manufacturer's instructions.
A leg of each adult moth and/or a portion of larvae collected were cut with a sterile blade and placed in a 2 mL tube. Insect genomic DNA was extracted using Isolate II genomic DNA extraction kit (Bioline, London, UK) as per the manufacturer's protocol. The resultant DNA was eluted in 50 µL Elution buffer (Bioline, London, UK) and quantified using a NanoDrop 2000/2000 c spectrophotometer (Thermo Fisher Scientific, Wilmington, NC, USA). Insect and plant DNA samples were stored at −20 • C for further downstream processing.

PCR for Plant Samples
General plant primers 3F_KIM_F/1R_KIM_R (3F_KIM_F 5 CGTACAGTACTTTTGT-GTTTACGAG 3 ; 1R_KIM_R5 ACCCAGTCCATCTGGAAATCTTGGTTC 3 ) were used to amplify a 900 bp region of the matK gene for the identification of host plants. Protocols for PCR of plant samples were similar to the PCR protocol for insect samples (Section 2.4.2), except the annealing step which was done at 49 • C for 45 s.

Agarose Gel Electrophoresis, PCR Product Purification and Sequencing
Resolution of the PCR product was done with 1% agarose gel stained with ethidium bromide (10 mg/mL) at 80 volts for 1 h (Bio-Rad model 200/2-0 power supply, Bio-Rad laboratories Inc., Hercules, CA, USA). DNA bands were visualized using an ultraviolet transilluminator and photographed using the KETA GL imaging system software (Wealtec Corp., Sparks, NV, USA). The resultant PCR products for both the insects and the host plants were purified using QIAquick PCR purification kit (Qiagen, Hilden, Germany) and

Sequence Analysis
Both plant and insect sequences were assembled and edited using Bioedit software v. 7.0.5.2 [44]. A consensus sequence generated from both the forward and reverse strand was queried on Basic Local Alignment Search Tool (BLAST) [45] and Barcode Of Life Data system (BOLD) [46] to determine similarity with sequences in the database. The default Species level barcode records were used. The top published hit on Bold was used for identification. Multiple sequence alignments were created on Clustal W [47]. Pairwise distances were generated using using Mega X [48]. Sequences were submitted to the GenBank (https://www.ncbi.nlm.nih.gov/WebSub/) (accessed on 10 March 2020) (see Table S2).

Occurrence Data
The occurrence data (species name, GPS co-ordinates) for the two species was collected during field surveys in Kenya and from the Global Biodiversity Information Facility (GBIF). A total of 96 points ( Figure 1a) were acquired (59 from field surveys and 37 from GBIF) for B. alcinoe, while the C. forda dataset comprised 70 points (57 from field surveys and 13 from GBIF) (Figure 1b).     [53]. RCP8.5 scenario predicts the mean global temperature increase projections of up to 3.7 • C. A collinearity test was conducted on the 19 bioclimatic variables to reduce collinearity between variables, to avoid overfitting of the model and variable inflation [54]. The variance inflation factor (VIF) test was used to assess the correlation between variables. The "vifcor" function in R software version 3.0.1 was used to run the VIF test [55]. Six bioclimatic variables, namely Bio2 (mean diurnal temperature range), Bio3 (Isothermality), Bio5 (max temp of warmest month), Bio13 (precipitation of wettest month), Bio15 (precipitation seasonality), and Bio19 (precipitation of coldest quarter) were selected for the B. alcinoe species analysis, while seven bioclimatic variables namely, Bio2, Bio4 (temperature seasonality), Bio8 (mean temperature of wettest quarter), Bio13, Bio15, Bio18 (precipitation of warmest quarter), and Bio19 were selected for the C. forda species analysis using a cutoff of |r2| > 0.7. Aside from the spatial correlation, the ecological relevance of the variables was also considered.

Model Calibration and Accuracy Assessment
Ecological niches of the two species were modeled using Maximum Entropy (MaxEnt) in the MaxEnt tool package version 3.4.1k which performs well for modeling presence only data [55]. The ENMEval package in R software was used to determine the required parameter settings to be used in Maxent software for the optimum tuning of the models [56]. Following the parameter settings from ENMEvaluate, three features (linear, quadratic and hinge) were utilized with a regularization multiplier of 3. The model calibration created the optimal models for the two saturniid species. The models were replicated 3 times using cross-validation method and an ensemble of the three probability outputs were used to determine the optimum suitability and performance of the models. Seventy percent of the presence records were utilized to train the model while 30% of the points were used to validate the performance of the model. The comparative relevance of each environmental variable for the models of C. forda and B. alcinoe was evaluated using the overall percentage contribution, area under the curve (AUC), and the Jackknife test. AUC values of 0 indicate impossible occurrence area while 1 indicates optimal occurrence area. The ROC method has shown to be effective in evaluating model performance and being independent of prevalence [57,58]. Outputs of the models highlighting the intensity and extent of habitat suitability of the two species were mapped with values ranging from 0 (unsuitable) to 1 (optimum). Suitability levels were grouped into five categories as follows: very low (0-0.1), low (0.1-0.3), moderate (0.3-0.5), high (0.5-0.7), and very high (0.7-1).

Morphological Identification of Edible Saturniids
Seven species of Saturniidae were identified in Kenya. They include B. alcinoe, C. forda, Gonimbrasia zambesina (Walker), Go. cocaulti Darge and Terral, Go. belina Westwood, Go. krucki (Hering) and Gynanisa nigra Bouvier. Dead larval stages of Gonimbrasia belina collected from Botswana and a sample of B. alcinoe collected from Nigeria were included for comparison (Table 1).

Bunaea alcinoe
Larvae are black with orange spots on the spiracles along the sides of the body. One B. alcinoe larva collected from Nigeria is red in color. Larvae of both color forms of B. alcinoe have white/yellow spines ( Figure 2). Bunaea alcinoe moths are dark brown in color with a large glass spot on the forewing ( Table 1). The hind wing has an orange eyespot ringed with black followed by white ( Figure 3).

Cirina forda
Larvae are black with yellow bands and white hairy spines. Cirina forda moths a smaller than the other moths. Cirina forda larvae in the same colony also presented in tw color forms. Some had a black body with yellow bands while others had a black body wi white bands ( Figure 2). It is light brown in color with a small black eyespot on th hindwing (Table 1, Figures 2 and 3).

Gonimbrasia cocaulti
Larvae appear black with whitish speckles and yellow spines. Cirina forda moths a smaller than the other moths. It is light brown in color with a small black eyespot on th hindwing ( Figure 3). Gonimbrasia cocaulti moths have a brownish ground color with ey

Cirina forda
Larvae are black with yellow bands and white hairy spines. Cirina forda moths are smaller than the other moths. Cirina forda larvae in the same colony also presented in two color forms. Some had a black body with yellow bands while others had a black body with white bands (Figure 2). It is light brown in color with a small black eyespot on the hindwing (Table 1, Figures 2 and 3).

Gonimbrasia cocaulti
Larvae appear black with whitish speckles and yellow spines. Cirina forda moths are smaller than the other moths. It is light brown in color with a small black eyespot on the hindwing ( Figure 3). Gonimbrasia cocaulti moths have a brownish ground color with eyespots on both the forewing and hindwing. The eyespot on the hindwing is white surrounded by reddish, black, and white rings while the one on the forewing is whitish circled by a brownish and white ring (Table 1, Figures 2 and 3).

Gonimbrasia krucki
Larvae are black with greenish-yellow speckles and black thick spines. Gonimbrasia krucki presented two color forms of their larvae. Larvae produced from the same egg clutch developed into forms that had a black body with either yellow or green speckles ( Figure 2). Gonimbrasia krucki moths have a yellow ground color with defined eyespots on both the forewing and the hindwing. Both eyespots are yellow in color ringed with black, pink, and red ( Figure 3).

Gonimbrasia belina
Larvae are red, grey, and green with black spines. Gonimbrasia belina moths are reddish brown in color with a brown eyespot on the hind wing circled by black and white rings. It has a small glass spot on the forewing. The front part of the hindwing is reddish in color (Figures 2 and 3).

Gonimbrasia zambesina
Larvae are black with yellow and grey speckles while some have black spines and others red spines (Table 1, Figure 2). Gonimbrasia zambesina moths occur in two color forms, green and brown. The green form is reddish purplish on the forward part of the hindwing. The eyespot on the hindwing is greenish yellow in the middle, circled by a black ring followed by greenish-yellow and white. The brown form has a yellowish-brown eyespot on the hind wing with black, pink, and whitish rings ( Figure 3). The brown form is not available in the NMK collection. The dichotomous key [12] reported the specimen as Go. said (Oberthuer). However, the author expressed uncertainty and suggested that it could be a form of G. zambesina. The green forms were collected from Kilifi, Embu and Kwale, while a mixture of the green and brown forms was collected from Makuyu in Murang'a County (all Kenyan sites). The green moths from Kwale and Kilifi produced larvae that were black with grey and yellow speckles and black spines ( Figure 2). Green moths from Embu laid eggs that hatched into larvae that were black with grey and yellow speckles and with red spines. The green and brown moths were also observed to mate with each other. The brown and green moths collected in Makuyu mated among themselves (green and green/green and brown/brown and brown) to produce black larvae with grey and yellow speckles and with red spines (Figure 3).

Gynanisa nigra
Larvae of Gy. nigra are green with white speckles and white spines ( Figure 2). We could not get adult moth from field collected Gy. nigra due to extensive parasitism.
Gonimbrasia krucki and B. alcinoe-Nigeria species had a 100% similarity to sequences from BOLD database. Gonimbrasia zambesina had a within species pairwise distance range of 0-0.61 and a range of 0-0.91 between species and BOLD sequence SAPBA773-07 (BOLD:AAD1339) ( Table 3). Gonimbrasia belina collected in Botswana and Gynanisa westwoodi had a 100% similarity within species, while they had a pairwise distance range of 0.15-0. 16 (Table 3). Gonimbrasia belina collected in Kenya was 100% similar to the BOLD sequence, while G. nigra had a pairwise distance of 0.61. Cirina forda showed a pairwise distance range of 0.3-2.13 within species and 0.46-2.13 between species and BOLD sequence SATWA281-07 (BOLD:AAB6982) ( Table 3).  Although Go. zambesina larvae and moths depicted different color forms morphologically, they identified as the same species using molecular characteristics. The genetic distance between all the Go. zambesina samples and the reference Go. zambesina sequence from BOLD (SAPBA773-07; BIN BOLD:AAD1339) was 0-0.61% while the genetic distance between all the samples and Go. zambesina sequence from BOLD (SAPBA772-07) was 1.07-1.23%. The genetic distance between the samples was 0-0.61%. All the samples included in the analysis had 658 bp (Figure 4).   (15) 0-0.

Molecular Differences among the Color Forms of Gonimbrasia zambesina and Bunaea alcinoe
Although Go. zambesina larvae and moths depicted different color forms morphologically, they identified as the same species using molecular characteristics. The genetic distance between all the Go. zambesina samples and the reference Go. zambesina sequence from BOLD (SAPBA773-07; BIN BOLD:AAD1339) was 0-0.61% while the genetic distance between all the samples and Go. zambesina sequence from BOLD (SAPBA772-07) was 1.07-1.23%. The genetic distance between the samples was 0-0.61%. All the samples included in the analysis had 658 bp (Figure 4).  The morphological difference of the two larvae color forms of B. alcinoe was also supported by molecular characterization. The red form with white spines (Nigeria-1) was 100% similar to SATWA891-07-COI-5P (B. alcinoe; BIN BOLD:AAA6756) from Burkina Faso. However, the same BOLD sequence had a 3.28-3.76% genetic distance from all the other black larvae forms of B. alcinoe collected from Kenya (see Table S4. All the other black forms with white spines from Kenya showed a genetic distance of 0.61-0.76% from LSAFR2238-12 (B. alcinoe; BIN BOLD:AAA6757) from South Africa. The same BOLD sequence had a genetic distance of 3.76% from the red color form collected from Nigeria. All the black forms clustered together with LSAFR2238-12 (B. alcinoe) while the red form clustered with SATWA891-07-COI-5P (B. alcinoe). All the samples included in the analysis had 658 bp ( Figure 5). The morphological difference of the two larvae color forms of B. alcinoe was also supported by molecular characterization. The red form with white spines (Nigeria-1) was 100% similar to SATWA891-07-COI-5P (B. alcinoe; BIN BOLD:AAA6756) from Burkina Faso. However, the same BOLD sequence had a 3.28-3.76% genetic distance from all the other black larvae forms of B. alcinoe collected from Kenya (see Table S4. All the other black forms with white spines from Kenya showed a genetic distance of 0.61-0.76% from LSAFR2238-12 (B. alcinoe; BIN BOLD:AAA6757) from South Africa. The same BOLD sequence had a genetic distance of 3.76% from the red color form collected from Nigeria. All the black forms clustered together with LSAFR2238-12 (B. alcinoe) while the red form clustered with SATWA891-07-COI-5P (B. alcinoe). All the samples included in the analysis had 658 bp ( Figure 5).

Distribution and Seasonality of Edible Saturniids in Kenya
The distribution of the edible saturniids in the various Counties in Kenyan is presented in Table S1. Gonimbrasia zambesina, C. forda, Go. krucki, and B. alcinoe were bivoltine occurring between April-June and October-December, reflecting the major and minor rainy seasons in the region. On the other hand, Go. cocaulti was univoltine and occurred only during the April-June season ( Table 4). The distribution of these saturniids was attributed to the availability of their host plants. The most widespread saturniid was B. alcinoe while the least widespread were Go. belina and Go. krucki which were only found in Kwale and Nairobi, respectively (Table 4).

Distribution and Seasonality of Edible Saturniids in Kenya
The distribution of the edible saturniids in the various Counties in Kenyan is presented in Table S1. Gonimbrasia zambesina, C. forda, Go. krucki, and B. alcinoe were bivoltine occurring between April-June and October-December, reflecting the major and minor rainy seasons in the region. On the other hand, Go. cocaulti was univoltine and occurred only during the April-June season ( Table 4). The distribution of these saturniids was attributed to the availability of their host plants. The most widespread saturniid was B. alcinoe while the least widespread were Go. belina and Go. krucki which were only found in Kwale and Nairobi, respectively (Table 4).

Area under Curve (AUC) Values
All the models using the current and future (RCP8.5:2050) showed a balance between goodness-of-fit and complexity (AUC > 0.80), for all the test and training datasets (Table 5). This demonstrates that our models showed good predictive performance.

Visualization of Habitat Suitability under Current and Future Climatic Conditions
Habitat suitability maps show that the tropics are optimal for both B. alcinoe ( Figure 6) and C. forda (Figure 7). Parts of the subtropical region in southern Africa are marginally suitable for B. alcinoe. For both MaxENT and GARP maps, a slight reduction in habitat suitability for both saturniids is predicted in future climate scenarios. Northern Africa is unsuitable in both present and future scenarios for B. alcinoe and C. forda.

Area under Curve (AUC) Values
All the models using the current and future (RCP8.5:2050) showed a balance between goodness-of-fit and complexity (AUC > 0.80), for all the test and training datasets ( Table  5). This demonstrates that our models showed good predictive performance. Habitat suitability maps show that the tropics are optimal for both B. alcinoe ( Figure  6) and C. forda (Figure 7). Parts of the subtropical region in southern Africa are marginally suitable for B. alcinoe. For both MaxENT and GARP maps, a slight reduction in habitat suitability for both saturniids is predicted in future climate scenarios. Northern Africa is unsuitable in both present and future scenarios for B. alcinoe and C. forda. a c

Discussion
This study has documented seven species of edible saturniids in Kenya. Three saturniid species, i.e., C. forda, Go. zambesina and B. alcinoe, are consumed in Kenya, mainly along the coastal belt along the Giriama community. Moreover, C. forda is widely consumed in West, Central and southern Africa [15,16,33,51]. Bunaea alcinoe is also a popular edible insect in West and Central Africa, for instance in countries like DR Congo, Cameroon and Nigeria [59][60][61], while Go. zambesina is highly popular in southern and Central Africa [17,59,62]. Gonimbrasia krucki is widely consumed in DR Congo [17], but not in Kenya. For Go. cocaulti, no records of human consumption are available from Kenya or elsewhere in Africa. However, due to the similarity of Go. cocaulti larva to that Go. belina, it is likely misidentified, given that they have been observed in consignments of mopane caterpillar in the UK [63].
Bunaea alcinoe, though black in colour and commonly found in Kenya, its genetic configuration is different from the red forms collected in Nigeria [64]. In Nigeria and DR Congo, both color forms have been reported feeding on the same host plant [17,64], yet this is the first time their genetic difference has been assessed. Further detailed studies relating the morphological and molecular differences, mating compatibility between color forms of B. alcinoe can shed light on their taxonomic status.
The sampled G. zambesina moths also exhibited two color forms and juveniles of the green adults carried black spines, as previously reported [12] and in http://www. africanmoths.com/ (accessed on 26 September 2019) [41]. The brown moths emerged from red spined larvae, which had been previously described as G. said [12], but inconclusively. Identification of edible saturniid species is important for the purpose of conservation [37,65], maintenance of quality in production [63] as well as mainstreaming consumption of the caterpillars.
We observed a bivoltine lifecycle in C. forda, with larvae occurring in April-June and October-December. In contrast, C. forda has been recorded as univoltine in Togo and Nigeria with larval occurrence between July and September [16,29] and in DR Congo with larvae appearing between November and January [17]. In all these cases, the occurrence of C. forda larvae coincides with the rainy seasons. For, B. alcinoe we noted a bivoltine lifecycle with larval appearance between April-May and October-December. In contrast, the same species in DR Congo is univoltine and occurs between October and May [17]. Understanding the temporal distribution of edible saturniids informs the need for mass production to ensure a continuous source throughout the year.
The model for habitat suitability of B. alcinoe and C. forda demonstrates that the two species thrive well within the tropical regions of Africa. However, B alcinoe spreads slightly into the subtropics, specifically in southern Africa. The model concurs with previous reports of availability and consumption of the two edible saturniids in southern, central and western Africa [11,16,17,28,33,51,59,66]. However, the availability of C. forda in the southern African region is not concurrent with previous reports [14,23] which recorded a wide distribution in the southern Africa region. This could be due to the limited data on the presence of the two saturniid species in the GBIF database which was used in this study. Future predictions for both species show a slight reduction in habitat suitability, stressing the need to conserve edible saturniid species' habitats.
The saturniids identified fed on specific host plants and consequently their availability depends on the occurrence of these host plants. We found B. alcinoe feeding on B. aegyptiaca and B. glabra similar to reports in Nigeria [63]. However, other research suggests a wide range of host plants, e.g., for DR Congo with Sarcocephalus latifolius (JE Sm.) EA Bruce  [63]. In our study in Kenya we observed, for the first time, larvae of C. forda feeding on E. divinorum, A. mearnsii and Manilkara sulcate, while in DR Congo it feeds mainly on Crossopteryx febrifuga [17]. In West Africa, C. forda is confined on the shea butter tree, Vittelaria paradoxa [16,29,33], whereas in southern Africa host plants include Burkea africana Hook (Fabaceae) and Albizia versicolor Welw. ex Oliv. (Fabaceae) [23].
We collected Go. zambesina from mango and cashew nut trees in Kenya, corroborating earlier findings [12]. Cashew nut and mango trees are important commercial trees whose nuts and fruits, respectively, are widely consumed. Gonimbrasia zambesina is sometimes considered a pest of mango trees [67]. Spraying of mango trees to curb pests may pose a threat to Go. zambesina larvae feeding on the leaves.
Apart from being host plants for edible insects, most of these plants have other uses in communities in Kenya and beyond. For example, E. divinorum is utilized by the Maasai in Kenya as firewood, their stem cuttings are used as toothbrushes and their fruits are edible [68]. Marakwets from the Rift Valley region in Kenya use E. divinorum as antivenom [69] while the Luo from western Kenya use it to treat venereal diseases [70]. Maasai also use V. tortilis and V. nilotica for firewood [71] while the Marakwet employ them for treating abdominal pains [69]. Balanites spp. are used to treat coughs [70], and finally A. mearnsii is often planted for firewood, timber, apiculture and a source of tanning dyes, and trees are also used for shade, nitrogen fixation and controlling soil erosion [72]. Such traditional knowledge can be used to encourage communities to conserve these plants and hence protect habitats for edible saturniid species.

Conclusions
We successfully documented seven species of saturniids in Kenya, among which three are consumed. The identity of these species was confirmed both at molecular and morphological level. Their distribution, seasonality and host plants were also established. We emphasize the importance of combining molecular barcoding, morphological identification, phenology, and ecology studies in identification of edible saturniid species. Potential habitats under current and future climate scenarios of two edible saturniid species, B. alcinoe and C. forda, were mapped. This information may help in implementing conservation measures for edible saturniids and their host plants. Due to their seasonal occurrence, further research is required on prospects for mass production to ensure a continuous supply and to prevent overharvesting from the wild forest for enhance sustainability. Moreover, potential economic benefits of edible saturniids for local communities in East Africa need to be quantified and their value chains established.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/insects12070600/s1, Table S1. Location and agro-ecological zones of the study sites in Kenya; Table S2. Accession numbers of saturniid sequences submitted to the GenBank. Table S3. Pairwise genetic distances of Gonimbrasia zambesina samples. Table S4. Pairwise genetic distances of Bunaea alcinoe samples.