Average values of proximate, mineral, and vitamin analyses are presented in Tables 1-3, respectively. Where applicable, standard deviations and the stage of development are shown. The MANOVA indicated that species had a significant effect (Pillai's trace F = 9.424, P<0.001) on all analyzed nutrients, except for Mo.
The three cockroach species differed distinctively in chemical composition. Six-spotted roaches contained the highest DM content (∼40–50%) of the three species. Stage of development significantly affected most nutrients (Pillai's trace F = 5.129, P = 0.007), as well as DM content, with the exception of Fe, Mo, and S.
All roaches contained high concentrations of crude protein (38–76% DM basis; Table 1), similar to values found in literature on American cockroaches (Periplaneta americana; 54% DM) [Bernard and Allen, 1997].
Cockroaches contained moderate-to-high concentrations of crude fat. In the earlier stages of development (small vs. large or adult stages), they contained more protein and less fat than larger specimens of the same species, as is true for most animals (a notable exception being neonatal rodents). Crude fat percentage increased with age in B. lateralis (from 14 to 27% DM) and G. portentosa (from 20 to 25% DM), but that same pattern was not present in E. distanti. The reported fat content of American cockroaches (28.4%) was slightly higher than the first two species, but lower than E. distanti [31–54% DM; Bernard and Allen, 1997].
In terms of dietary “fiber” content, both neutral detergent fiber (NDF) and acid detergent fiber (ADF) content were similar for B. lateralis, averaging about 12% of DM. Approximately 60–90% of ADF in insects is chitin provided by the exoskeleton [Barker et al., 1998; Finke, 2007; Oyarzun et al., 1996]. The ADF content of G. portentosa was 10–13% of DM. However, NDF in this species was considerably higher (∼36% of DM) and may represent true dietary fiber from vegetables in the digestive tract. Both body and gut content, especially in species with a relatively large gut or consuming high fiber diets, contribute to the nutrient content of feeder prey species. Thus, diet may provide essential nutrients otherwise unavailable from the insect with an empty gut [Finke, 2003; Klasing et al., 2000].
Total ash content of E. distanti was significantly lower (2–4% DM) than in B. lateralis (7–8% DM; P<0.001)) and G. portentosa (4–8% DM; P = 0.007), similar to American cockroaches [3.3% DM; Bernard and Allen, 1997]. Mineral content among the three cockroach species differed greatly (Table 2). As expected [Barker et al., 1998; Finke, 2002; Studier and Sevick, 1992], an inverse Ca:P ratio was found in cockroaches. Therefore, if using cockroaches as a feeder species, Ca supplementation is necessary to achieve a Ca:P of 1:1 [Donoghue and Langenberg, 1994]. Larger invertebrates (adult or large nymphs) contained lower concentrations of most minerals (Ca, P, Mg, K, Na, Zn, Cu, Mn, and Mo) compared with smaller sized individuals of the same species. Nonetheless, roaches seem to be an excellent dietary source of Zn and Cu. Fe content in E. distanti and G. portentosa increased with age. Because excess dietary Fe can contribute to Fe storage diseases in several species of birds and mammals [Bonar et al., 2006; Farina et al., 2005; Sheppard and Dierenfeld, 2002; Williams et al., 2008], it is important to know all contributory factors for Fe intake.
Vitamin E content of cockroaches was relatively low (11–16 mg/kg DM; Table 3), providing approximately 20 IU vitamin E/kg DM (1 mg = 1.49 IU). Pennino et al.  found almost 10-fold higher concentrations of vitamin E (179 IU/kg DM) in wild-caught cockroaches. Retinol content varied from 25 to 116 mg/kg DM; therefore, calculated vitamin A activity (0.3 µg retinol = 1 IU) was low (<400 IU/kg DM) compared with estimated requirements, using domestic felids as a carnivore model for insectivores (∼5,000 IU/kg DM maintenance; 9,000 IU vitamin A/kg DM, growth, and reproduction; [NRC, 2006]. As with vitamin E, free-ranging cockroaches reported by Pennino et al.  contained considerably more vitamin A (1,000 IU/kg DM) than the cockroaches in this study. Lutein, zeaxanthin, and β-carotene was found in all samples. Although dehydrolutein (DHL) and anhydrolutein (AHL) are metabolites of lutein, DHL was not quantifiable in B. lateralis or E. distanti, and AHL was only found in E. distanti samples. Both β-carotene (Bcar), found in all three cockroach species, and β-cryptoxanthin (Bcry) have provitamin A activity in many species [McGraw et al., 2006]. Owing to the widely varying molecular structures of carotenoids, there might be species-dependant differences in the ability of vitamin A synthesis from these compounds. Because vitamin A deficiency has been reported for insectivores fed unsupplemented invertebrates [Ferguson et al., 1996], vitamin A metabolism could be explored among different cockroach species fed identical diets to evaluate synthetic pathways, and determine optimal dietary regimens/ingredients for production of feeder insects with the most appropriate vitamin A levels.
Compared with mealworm and superworm larvae, rusty red roaches (B. lateralis) and hissing cockroaches (G. portentosa) provide high protein, lower fat alternative food items for insectivores—more similar to cricket proximate nutrient composition [Barker et al., 1998; Bernard and Allen, 1997; Finke, 2002; Jansen and Nijboer, 2003; Oonincx et al., 2010; Pennino et al., 1991]. Six-spotted cockroach nymphs (E. distanti), on the other hand, tended to be higher in fat and may be a poorer source of protein than either the other roach species, crickets, or beetle larvae. Owing to their high fat content, they may be considered a high-calorie treat item or could prove useful for improving body condition of insectivores. Mineral content of roaches was variable, depending on species, size, and diet, but all roaches examined still contained inverse Ca:P ratios, in the same ranges as the more commonly fed invertebrate prey [Barker et al., 1998; Bernard and Allen, 1997; Finke, 2002; Jansen and Nijboer, 2003; Oonincx et al., 2010]. Other macrominerals were found in concentrations that would be considered adequate to meet known nutritional requirements of domestic felids [NRC, 2006], considered to be the most suitable physiologic model for insectivores. Conversely, some microminerals, particularly Fe, could be excessive. Small hissing cockroaches are similar in body size to adult house crickets, and may provide a suitable nutritional substitute for crickets in insectivore diets (if consumed by the insectivore).
D. melanogaster samples contained high levels of crude protein along with moderate levels of crude fat, dietary fiber, and ash (Table 1). Ca:P ratios were imbalanced (0.13:1), but D. melanogaster seems to provide adequate levels of other macro-and microminerals measured; Fe and Zn concentrations were particularly high. In this and in Barker et al.'s study , the same commercial diet was used, which may underlie the high Fe content as well as the higher fat content compared with the other published values [18–19 vs. 13% of DM; Table 1; Jansen and Nijboer, 2003; Barker et al., 1998; Bernard and Allen, 1997; Finke, 2002].
In contrast, vitamin E in this study (112 mg/kg DM or 166 IU/kg DM) was considerably higher than levels reported by Barker et al. [1998; 23 IU/kg DM]. Only small traces of lutein, β-cryptoxanthin, β-carotene, and vitamin A were found in D. melanogaster. Barker et al.  reported undetectable levels of vitamin A in fruit flies. Insectivores eating a relatively large proportion of D. melanogaster might benefit from dietary vitamin A and/or carotenoid supplementation.
Regarding size, D. melanogaster may be an alternative size option for species that consume pinhead crickets (A. domesticus). Fruit flies are similar in protein content as pinhead crickets (68 vs. 55–68% of DM, respectively) and have a higher crude fat and vitamin E content [Barker et al., 1998; Bernard and Allen, 1997; Finke, 2002]. Ash content is similar between the two insects (7 vs. 5–9% DM), but there are compositional differences in specific minerals which may be of consequence (i.e. Na (0.31 vs. 0.43–0.59% DM), Fe (401 vs. 93–200 mg/kg DM), and Mg (17 vs. 30–39 mg/kg DM). Additionally, vitamin E found in D. melanogaster differed greatly from that reported in pinhead crickets [167 vs. 40–70 IU/kg DM; Finke, 2002; Barker et al., 1998].
Native M. rhombifolium contained the highest percentage of crude protein, the lowest crude fat, and the highest dietary fiber content of all invertebrates evaluated in this study (Table 1). High concentrations of ash and most minerals are noted, but the Ca:P ratio was still low (0.27:1), and Na was low compared with the other investigated species. M. rhombifolium also contained very high levels of vitamin E and carotenoids (with the exception of DHL; Table 3). Vitamin A content was high compared with the other investigated species and literature data, except for silkworms [Barker et al., 1998; Finke, 2002]. These high ash and vitamin values are likely owing to the diet of fresh green plant materials consumed, as was observed with other herbivorous insects [Dierenfeld, 2002; Dierenfeld and Fidgett, 2003]. Therefore, M. rhombifolium may be an important source of carotenoid pigments, fat-soluble vitamins, and minerals for insectivorous species. However, high fiber content may limit palatability and/or bioavailability of some nutrients.
P. scaber was strikingly different in composition compared with all other invertebrates analyzed in this study. Protein (41% DM), crude fat (12% DM), and fiber (<15% DM) concentrations were relatively low compared with data from other studies on woodlice species [protein 40–80% DM; Pokarzhevskii et al., 2003]. The ash content of woodlice was exceptionally high (33%). As a terrestrial crustacean, woodlice have a mineralized exoskeleton and contain high levels of Ca, Mg, Na, Fe, and Cu (Table 2). The exoskeleton can contain up to 24% Ca [Becker et al., 2005]; similar concentrations (10–15% Ca) have been described for other woodlice species [Bureš and Weidinger, 2003; Graveland and Vangijzen, 1994; Pokarzhevskii et al., 2003; Reichle et al., 1969]. The calculated Ca:P ratio of 12:1 is high; this species can possibly be used to correct dietary Ca deficiencies in insectivore diets. Bioavailability and palatability studies are recommended. Woodlice exoskeletons, however, were not found in the feces of two lizard species (Pogona vitticeps and Diploglossus warreni) during woodlice acceptability trials (personal observation).
For the isopod Armadillum vulgarum, concentrations of Fe (3,170–3,390 mg/kg DM), Zn 332–341 (mg/kg DM), Cu (347–410 mg/kg DM), and Mn (15–16 mg/kg DM) are reported [Peters et al., 2005]. In the same species, a high Mo concentration (0.3 mg/kg DM) has been reported, which is more than threefold the concentration of P. scaber in this study [Anke et al., 2007]. Whether these differences are owing to species, diet, or substrate is unknown. Nonetheless, it is apparent that mineral concentrations in woodlice are variable and highly dependent on soil conditions. For example, the Zn concentration in Oniscus asellus varied between 54 and 499 mg/kg DM, in which the extremely high concentration was found in specimens originating from a Zn mine [Hopkin and Martin, 1982]; Cu concentration varied from 82.1 to 163 mg/kg DM in the same study.
The concentration of K in isopods measured in this study (0.93% DM) agrees with reported concentrations for other species of woodlice 0.87–1.32% DM) [Pokarzhevskii et al., 2003; Reichle et al., 1969]. P. scaber contained a fairly high concentration of Na (0.81% DM), similar to the woodlouse Ligidium blueridgensis (0.72% DM) [Reichle et al., 1969]. For the woodlouse Philoscia muscorum, a concentration of 0.19% Mg DM was reported [Pokarzhevskii et al., 2003], which is much lower than P. scaber in this study (0.47% DM).
Tenebrio beetles and Zophobas beetles contained more protein than literature values reported for their larval counterparts (51 and 43% DM, respectively). This might be associated with a higher degree of sclerotization in beetles, which could negatively influence digestibility [Finke, 2007]. Additionally, the Tenebrio and Zophobas beetles contained one-half to one-third of the fat content of their larval counterparts, respectively, as well as a higher ash content (Table 1) [Barker et al., 1998; Bernard and Allen, 1997; Finke, 2002]. In general, macromineral concentrations in beetles were similar to, or slightly lower than, published values for larval stages of these beetles [Table 2; Barker et al., 1998; Bernard and Allen, 1997; Finke, 2002], whereas trace mineral concentrations in the beetles studied were consistently higher than previously reported values for both larval and adult stages. These data provide the first published information on mineral content of adult Z. morio. Furthermore, feeding trials conducted separately indicate beetles are readily accepted by certain lizard species in captivity (Oonincx and Dierenfeld, unpublished).
The concentration of the measured fat-soluble nutrients was low in the beetles. However, large differences were apparent.
The Zophobas beetles contained twice the vitamin E, 3.5 times the concentrations of vitamin A and β-carotene, and 10 times more lutein compared with Tenebrio beetles, and thus may provide a more significant source of these nutrients. It might be interesting to evaluate these nutrients further in all life stages of the two species, to better understand metabolic pathways and ultimate storage as a source of these particular nutrients for insectivores.