Insect diets as mixtures: optimization for a polyphagous weevil

Authors: Lapointe, Stephen; Evens, Terence; Niedz, Randall

Submitted to: Journal of Insect Physiology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/23/2008
Publication Date: 8/7/2008
Citation: Lapointe, S.L., Evens, T.J., Niedz, R.P. 2008. Insect diets as mixtures: optimization for a polyphagous weevil. Journal of Insect Physiology 54: 1157-1167.

Interpretive Summary: Development or improvement of artificial insect diets can be tedious, convoluted and under-appreciated. Artificial diets are often derived empirically and their complexity makes refinement difficult. The advent of modern hardware and software allows application of complex statistical methods to this problem. We applied a multidimensional design to optimize an artificial diet for Diaprepes abbreviatus, a highly polyphagous pest of citrus and other crops from the Caribbean and southern U.S. We identified components of the currently available commercial diet that are unnecessary for larval development. We also identified different diet blends that yield insects of greater weight, and other diets that produced larger numbers of adults compared with the commercial diet. A high cottonseed meal diet yielded the greatest survival of larvae to adult, the greatest biomass produced per unit of diet, and the lowest cost of production per adult insect. Many insect-rearing programs could benefit from this approach to optimizing complex diet mixtures.

Technical Abstract: Development or improvement of artificial insect diets can be tedious, convoluted and under-appreciated. Modern hardware and software can be applied to efficiently sample multidimensional designs to analyze diets as n-dimensional mixtures. We identified a set of response-optimized meridic diets that contain fewer ingredients than the current commercial diet for Diaprepes abbreviatus, a polyphagous weevil pest of the Caribbean and southern U.S. A diet blend optimized to produce maximum adult weight was predicted to produce adult D. abbreviatus that weigh 28% more compared with adults reared on the commercial diet. Blends that predict high individual adult weights resulted in low survival compared with those blends predicted to yield adults of more modest proportions. In contrast, a simplified high cottonseed meal blend produced smaller adults more similar to field-collected individuals, and produced the greatest number of adults and the greatest biomass at relatively low cost compared with diets that yielded adult weevils of high weight. We believe that many insect-rearing programs would benefit from application of these methods to situations where diet optimization is desired for researcher-selected criteria. This approach is broadly applicable to any problem that can be conceptualized as a mixture problem.