We Rear Insects

First, if you like what you see in today’s blog, you are bound to like how I will be teaching this approach to AI-based neural networks and Design of Experiments in my upcoming (May, 2026) Insect Rearing Fundamentals class.

I have been touting the advantages of using design of experiments (DoE) to help us better understand interactions of components in insect rearing systems and to further our grasp of which components are potent forces towards the improvement of our systems. In this blog entry, I am trying to explain how we might start some initial rearing system inquiry using the powerful AI tool neural networks (often called artificial neural networks). I use the JMP statistical package for my DoE and other statistical practices, and I have found the JMP procedures to be very user friendly for people like me who are NOT mathematicians or statisticians. This is especially the case for my learning to apply neural network processes to my rearing inquiries.

For one thing, in my many ventures into websites and other resources that help explain the nature, value and procedures in neural networks, I found a most compelling, entertaining, and clear tutorial from Professor Dmitry Shaltayev (https://youtu.be/u-ngF1YXqhY ). Though his presentation does not use examples from insect rearing, he is very clear and very compelling in his tutorial. First, he starts with the derivation of neural networks from the mammalian nervous system where neurons have dendrites, cell bodies, and axons that receive and process neurological information. He further explains how neural networking/AI is central to facial recognition, self-driving cars, and many other applications. He also uses the JMP software, so his explanations are very useful to me in my JMP neural network adventures.

Secondly, I must mention (by way of a plug for my book: Design, Operation, and Control of Insect Rearing Systems, Cohen. 2021. CRC Press. Boca Raton FL) that I have already made an argument for using neural networks, and my examples here are taken from my discussion.

In this study, I had used design of experiments to set up an experiment with diets for painted lady butterflies, using their weights as the responses and several diet components as the factors (or experimental variables). Here is the data table representing these values:

Next I fed the terms (factors and responses) into a JMP neural network analysis and got this diagram to represent the inquiry;

First, please note the Prediction Profiler (a typical output of JMP analyses) and the little graphs that show, for example, that between agar and gellan as gelling agents, agar use consistently resulted in greater biomass. Also, soy flour showed a slight but consistent tendency to yield higher body weights than wheat germ; brewer’s yeast here seemed slightly better than torula yeast, and casein gave much better results (higher weights) than did whey protein.

Second, I point out that this system of running multiple factors (rather than one factor at a time) is much more useful in understanding interactions of the components and giving us a much more realistic experimental framework where the factors are “dissected out” to show their contributions to the responses.

Third, I caution the readers that this set of experiments is still not definitive. The outcomes may vary according to which species are being explored, which sources of the components are being used (agar varieties differ from supplier to supplier and even from batch to batch).

I will address all this and more in near future blogs. I hope you tune in, and I hope you are encouraged to think about using DoE and neural networks to develop and improve your rearing systems!

Happy Rearing!

This mulberry silk moth (Bombyx mori) is a product of over 5,000 years of domestication and rearing. The domestication of the hundreds of species of insects that we currently rear should be a testimony that rearing is not for dummies. In contrast to the popular trend of simplifying everything from Artificial Intelligence to Vegetarian Cooking (over 300 “For Dummies” titles listed in Amazon), I am going against the grain in touting my writings and teachings (my online and in person classes) as being intellectually based and seeking to give a deep and far-reaching understanding of the complexities of insect rearing.

I hasten to say that I think it’s a clever way to approach topics that would put people off (Economics, Physics, Religious Philosophy, etc.), but without deliberately relegating insect rearing as a “for dummies” concept, the entomological community has done a disservice to the potential of insect rearing to contribute even more than it has to the well-being of humanity.

I have devoted several recent posts and will devote several more to make my point about how the statement attributed to Socrates “The unexamined life is not worth living” applies to the complex practice of insect rearing.

I hope you will examine some of my posts.

==Allen Carson Cohen

I continue here with an explanation of my approaches to insect rearing teaching and research. In my recent posts, I discussed my early efforts to develop a practical artificial diet for predatory insects, mainly Hemiptera/Heteroptera species. I confessed my many failures, disappointments, and frustrations that resulted from my (and the entomology community, in general’s) ignorance about the true feeding habits of the big-eyed bugs and other insects with piercing and sucking mouthparts. In the past few blogs, I mentioned that my realization that the predators were using extra-oral digestion (EOD) and that EOD had far-reaching implications for understanding feeding biology of many species of insects and other arthropods. In fact, right now (February of 2026), I am writing a paper on the discovery of EOD as a feeding “strategy” of the wood-eating buprestid beetle, the emerald ash borer! I also must point out that a fairly recent paper by Ramsey et al. 2019 (with myself as one of the co-authors) clearly documented that Varroa mites, which were long considered hemolymph feeders were indeed feeding on fat body and other semi-solid tissues and NOT hemolymph. They were using EOD to liquify the honeybee’s tissues. Case by case, when we learn in depth the feeding dynamics of many (most?) species of insects, we find that they are doing some sort of pre-oral/extra-oral processing of the food.

What results did we get from recognition and understanding of the true feeding strategy led to development of artificial diets for big-eyed bugs and other species, including green lacewing larvae (Chrysopidae). Despite this clear example of the benefits of “knowing our insects,” too often researchers (like me in my novice insect rearing researcher days) try to accomplish difficult goals without adequate knowledge of their subjects. In my writings and teachings, I try to illustrate the benefits of in depth, mechanistic knowledge of the insects and other rearing system components (diets, environment, microbial relations, containers, hidden interference in our rearing conditions) for success in developing rearing systems and in maintaining working systems. This takes us back to my main point here: exploring how and why, rather than what is the best path to rearing success!

More about HOW and WHY vs. WHAT or Rationale-Based Inquiry

When I write or speak about the efficacy of understanding HOW and WHY, I am getting at the idea that rationale is an all-important part of experimentation. When I would do science demonstrations with my wife’s 5th graders, they always wanted to see what would happen if…. This is a great thing to be curious, but advancing past simple curiosity, we do better (less randomly) if we have a knowledge-based reason for what we experiment with. To test my commitment to this concept, I used a word search in my Insect Diets book and found more than 70 uses of the term rationale, and in the Design, Operation… book, I found more than 100 times that I used rationale to explain how or why a material was used or a process was applied. For example, in both books, I tried to probe deeply into wheat germ to ask why it was so widely useful as a diet component (now for hundreds of species of insects and the production of trillions of insects). But many decades ago, the great Dr. Erma Vanderzant had the same type of curiosity about the wonders of wheat germ.

The excerpt below (Figure 5) shows the first page of a 1967 paper where wheat germ’s qualities are explored by Vanderzant: she explained many of the qualities of wheat germ. Clearly Vanderzant went back to her first use of wheat germ in 1959 considering why the wheat germ did not work well with boll weevils, but why it DID work well with pink boll worms and subsequently with various other insects. She discusses the proteins, lipids, carbohydrates, minerals, and B-vitamins and the other components that “worked” or did not “work” for various insects.

By the way, I had wondered what originally motivated Vanderzant to use wheat germ in her boll weevil and the 1960 (Adkisson et al.) diet for pink bollworms. I had even contacted (some time after her death) a member of her family to ask if he knew what her reasoning was when she first tried wheat germ, and he did not know.

This concludes today’s discussion of the benefits of rationale-based inquiry into the WHY’s and HOW’s of insect diets and other rearing systems components. I will continue this discussion as a basis for my zeal to influence rearing researchers to consider as many facets of their experimental factors as possible. Please watch for more posts to follow soon!

Figure 1 (Feb. 22, 2026 Blog). Euthyrhynchus floridanus (the Florida predatory stink bug) 2nd instar nymphs feeding on a Tenebrio molitor pupa. This colony came from the late Professor William S. Bowers from the University of Arizona. The large size and rather easy rearing of this predatory stink bug allowed us to use it as a model system to study extra-oral digestion (EOD) and other features of its hemipteran/heteropteran biology.

By December of 1983, I had finally learned that the original concept or model of feeding by predators from the sub-order Heteroptera did NOT make their living by sucking liquids (essentially hemolymph) from their prey, but instead, they used EOD to inject digestive enzymes that turned the prey’s internal contents from solid or semi-solid structures like muscles, fat body, etc. into a runny slurry of cellular debris and dissolved nutrients. In my early studies of EOD, I learned the more accurate model of feeding that allow the predators to use much more nutrient rich foods than simple hemolymph. For example, the ingested slurry was at least 5 to 10 times as concentrated with proteins, lipids, and other key nutrients than hemolymph itself.

I will address the various outcomes of using EOD by many, many species of arthropods, but for now let me turn to the practical issue of diet development. If our target predators were ingesting hemolymph, they were getting more than 90-95 mg of water from each 100 mg of ingested hemolymph (leading to hypothesis 1: strictly liquid feeding). However, if the predators were ingesting all (most) of the insides of their prey, they would be consuming about 50 mg of water for every 100 mg of ingested material (slurry). This means that through EOD, they were about 10 X more efficient at extracting solid, high-nutrient food from their prey. Therefore, our second hypothesis was that through EOD, the predators were feeding selectively and more efficiently on high nutrient materials. The affirmation of this 2nd hypothesis led to the “if…then…” conclusion that if we offered the predators non-insect diet components such as meat products, their food would be much more like insect insides in their nutrient concentration, texture/consistency, and digestibility. The experiments with meat diets proved successful after we made adjustments for proportions of various nutrients and a suitable presentation system that would adequately mimic the natural prey, leading to the below publication:

Figure 3 (Feb. 22, 2026 Blog)Showing two Geocoris punctipes feeding on the Cohen 1985 artificial diet made from beef liver, ground beef, and sucrose solution.

Figure 4 (Feb. 22, 2026 Blog) shows two Euthyrhynchus floridanus (predatory stink bugs, family Pentatomidae) feeding on the Cohen 1985 diet developed for G. punctipes (see Figure 3).

Now, let’s see where we are. The diet which I developed by learning how predatory Heteroptera actually feed, using extra-oral digestion, rather than the “drinking straw” concept was so successful that these predators and others such as the predatory stink bugs could feed on it and develop and grow effectively. This leads to the main point of these blog discussions: By knowing the mechanisms of our insects’ biology which we learn through careful observation supported by experimentation, we have much greater power over the systems we are working with.

Once I started to have some successes (among MANY more failures than successes), I became transfixed by the complexities and the rewards of insect rearing, and I devoted the rest of my professional life to discovery of insect biology through science-based inquiry into the insects themselves and the rearing systems we design for the insects. The drama and excitement of this kind of discovery and the applications that come from it are the basis of all my teaching and all the research I have been doing and hope to continue to do.

I’ll explore more of this in blogs on this site. I hope you will join these explanations and the suggestions that I offer in the website and in my courses.

Happy Rearing!

Allen Carson Cohen teaching a recent Fundamentals of Insect Rearing class (Figure 1 for Feb. 21, 2026 post).

SOME BIOGRAPHICAL ANECDOTES BEHIND MY APPROACH TO INSECT REARING EDUCATION. Hi insect rearing community! Though I have interacted with many of you from the rearing community as colleagues, students, and others, I decided to provide some background that explains my current involvement in insect rearing.

I started my formal professional work with insect rearing as an older post-doc at the University of Arizona and the USDA, Agricultural Research Service in 1979. Prior to accepting that position, I had been teaching biology at Cypress Community College, in Cypress, California for over a decade, and I did my doctorate in entomology at UC, Riverside during my Cypress teaching phase. The job at U of A and at the USDA was to develop an artificial diet for predators to be used in a mass-rearing program to control crop pests, mainly in cotton and alfalfa cropping systems. At the start of my post-doc, I knew nothing of insect rearing, as my background was with field-collected blister beetles (Family: Meloidae). Because I had some background in physiological and biochemical ecology of insects, Drs. George Ware and Harry Graham thought I might find my way through the complexities of diet development.

Figure 2 (Feb. 21, 2026 post). Big-eyed (Geocoris punctipes) bug eating a cotton aphid (Aphis gossipii) (picture taken by Jack Dykinga in Cohen’s USDA, Western Cotton Lab, Phoenix, an ARS lab, circa 1995). Please note that this image is part of the rearing course I deliver online (Zoom) several times a year.

The picture of the big-eyed bug eating an aphid typifies what I was seeing in the first years of my efforts to develop an artificial diet that would support a program of mass rearing of this or other predators. Seeing the piercing-sucking mouthparts of the big-eyed bug penetrating the aphid, I assumed (and was informed by the then current literature) that the hemipteran predator was inserting a kind of drinking straw into the prey and sucking out the fluids (hemolymph) of that prey.

For nearly 3.5 years, I struggled with efforts to design a liquid diet that would match the hemolymph chemistry of various prey of the hemipteran predators (beet army worms, corn earworms, aphids, etc.) I used gas chromatography, HPLC, amino acid analyzers (ion exchange systems), atomic absorption spectroscopy, etc. to analyze the contents of hemolymph and then I would try to formulate liquids that reflected the lipids, amino acids, carbohydrates, etc. of prey hemolymph, and I struggled further to develop a liquid diet presentation system that would simulate the cuticle and body configuration of typical prey. ALL THESE EFFORTS WERE OF NO AVAIL!

Figure 3 (Feb. 21, 2026 post) illustrates the system I used to provide a coated liquid diet for my big-eyed bugs (left side). It also shows an interim paper that I wrote as a progress (or lack of progress) report on liquid diets for G. puncitpes.

SUCCESS AT LAST! It was not until I started studying the true feeding habits of big-eyed bugs and their hemipteran (heteropteran) cohorts that I was able to apply the basic knowledge I gained about extra-oral digestion to develop a solid/semisolid artificial diet for G. puncitpes and, in turn, for other arthropods.

Figure 4 (Feb. 21, 2026 post) showing the strides that I had made in diet development (practical accomplishments) and the deeper understanding of extra-oral digestion (= EOD) in basic insect science (biochemistry, morphology, behavior, ecology, physiology). The 1998 paper (on the left) represents the publication on EOD, which I called “Solid-to-Liquid Feeding: The Inside(s) Story of Extra-Oral Digestion in Predaceous Arthropods. The images on the right side of Figure 4 show the kinds of extra-oral digestion (what I called non-refluxing and refluxing feeding), and the middle images show the structure of extracellular matrix that holds together the cells of the prey. The paper was a feature article in a 1998 American Entomologist.

will leave off here for today’s blog post, but I will take up from here with some specifics about how and why I developed my approaches to rearing education, including the reasoning behind rationale-driven, basic science-based inquiry. You will soon see how and why my emphasis on learning cause-and-effect relationships, rearing systems interactions, and mechanism-based inquiry offers a high likelihood of success in developing new rearing systems, improving (optimizing) existing systems, and developing data-driven standard operating procedures (SOPs) for rearing systems.

I have been a researcher in insect rearing since 1979. During the past 47 years, I have done hands-on research, writing, reviewing, reading, and editing rearing studies and related studies of all aspects of insect rearing. I have had many successes and many more failures in my efforts to develop or improve rearing systems. I have struggled with the entomology community and the infrastructure of rearing, and from my experiences with doing and thinking about rearing EVERY day for these 47 years ~17,000 days for maybe 8 hours a day–yes, I work on weekends, too–or about 135,000 hours of thinking about rearing research!!!, I am beginning to develop a wisdom and understanding of things that work and things that fail to work in our quest to improve the state of insect rearing.

CONFESSION: Though I said I was a rearing researcher for 47 years, I have also been a teacher for over 60 years (starting with high school English teaching which I began in 1965), and above my passion for rearing research, I have an even deeper passion for teaching–sharing whatever I know with everyone who cares to learn from me. Much of what I say here about rearing research is what I’m calling an Inquiry-based approach that has come from my ongoing teaching.

Research and Teaching: Just before I was discharged (honourably) from the US Marine Corps, I was having some dental work done at El Toro Marine Base, and the dentist asked me what I was going to do after my discharge and I proudly told him I was going to become a teacher. His brusque reply was, “Oh, everyone knows if you can’t DO, you TEACH.” I never forgot this insult, and to this day I attribute my ever-deepening understanding of rearing science to the combination of my teaching with my research. Every time I teach a subject or concept, I get to understand it more deeply.

Doing and Teaching/Teaching and Doing: The more I worked with students of insect rearing, the more I realised how under-supported rearing education was. In my own publications and the thousands of rearing publications in the literature, I saw the same pattern in reports: WHAT works, but little attention to HOW or WHY it works = RATIONALE. Vanderzant’s group (Adkisson et al. 1960) taught us that wheat germ works for pink bollworms (or later for tobacco hornworms and hundreds of other insect species), but they did not offer reasons behind trying wheat germ in insect diets.

Yamamoto 1969 taught us that torula yeast works in the diet for tobacco hornworms, but he never explained how or why it works. These two papers were revolutionary in the doors they opened for insect rearing with artificial diets, so learning what works is super important. But if we can understand how and why wheat germ and torula yeast work as they do, we have amplified power that comes with knowledge to make more likely that we will be able to predict or analyze failures and we have far more power to project how rearing components can be used for other purposes. One of my axioms of insect rearing is DETAILED RATIONALE EMPOWERS US TO INNOVATE. Erma Vanderzant (1967) herself returned to wheat germ to explain much of the characteristics/qualities of wheat germ to help spell out the basis of its value in insect diets. In her 1967 paper, Vanderzant provided a model for how researchers can focus on qualities of a diet component (wheat germ’s quality protein, unsaturated fatty acids, vitamins, etc.) that confer upon a material a deeper understanding of how and why it fits the needs of rearing practitioners.

Above is a photo of the late Dr. Erma Vanderzant taken from her obituary. In this tribute to Dr. Vanderzant, I tried to celebrate 50 years of success (1960-2010) with wheat germ in insect diets.

It is this HOW and WHY that lead to explicit rationale that drives my teaching about insect rearing. My courses are designed to treat extensively the mechanistic, cause and effect thinking that leads to treating insect rearing as a quantitative science, rather than an anecdotal art. I will discuss this topic further in near future blogs.

Tobacco hornworm neonates on Yamamoto Diet (top): the diet formulation as prescribed by Robert Yamamoto 1969 (bottom). This paper has been cited more than 300 times, and it served as the basis for a paper by Bell and Joachim 1976, cited well over 1,000 times. Both papers report highly successful use of wheat germ and torula yeast as major components of diets that have supported production of billions of hornworms over the past 50 years.

Adkisson, P.L., E.S. Vanderzant, D.L. Bull, and W.E. Allison.1960. A wheat germ medium for rearing the pink bollworm. Journal of Economic Entomology. 95: 256-260.

Bell, R.A., and F.G. Joachim. 1976. Techniques for rearing laboratory colonies of tobacco hornworms and pink bollworms. Annals of the Entomological Society of America: 69: 365-373.

Yamamoto, R.T. 1969. Mass rearing of tobacco hornworms II. Larval rearing and pupation. Journal of Economic Entomology. 62: 1427-1431.

Vanderzant, E.S. 1967. Wheat-germ diets for insects: rearing the boll weevil and the salt-marsh caterpillar. Annalsof the Entomological Society of America 60: 1062-1066..

Insect Rearing Fundamentals: May, 2026

This is a compressed version of the course syllabus. Please note that each day’s class meeting is 3 hours long, including a 5 minute break at the halfway point. I make frequent stops to ask for questions or other student/participant input. The course is designed to include materials from my books, Insect Diets: Science and Technology 2nd Edition 2015 AND Design, Operation, and Control of Insect Rearing Systems 2021, both from CRC Press. The live, real-time classes go beyond some aspects of the books where 1) materials are constantly updated as I become aware of new information, including my own original research and my activities as an Academic Editor of the MDPI journal, Insects.

As with recent offerings of the Fundamentals course, the registration fee is $450 (US$), and the information for registration is specified below in a previous blog page.

As always, I invite specific questions about the course through contact with me at this email address: accohen@ncsu.edu

In May of 2026, Allen Carson Cohen will be teaching an online insect rearing fundamentals course. The course is 24 hours of instruction via Zoom with lots of interaction between participants (students) and Professor Cohen. We have updated this course to include new information about rearing systems, their standard operating procedures, using statistically-based optimization, engineering of the rearing systems, and up-to-date technology for measuring quality and process control. While getting into new approaches, the course also contains standard information that has been covered conventionally in these courses and in Professor Cohen’s books and recent papers.

Here is an example from Allen Cohen’s recent research on interactions of components in diets, using a JMP statistical system’s format for diet texture using the 1969 Yamamoto Diet for tobacco hornworms (Manduca sexta).

In this figure, we can see an example of the kinds of exploration we can do with a design of experiments (DoE)-based inquiry. In the experiment illustrated here, we were asking the question: how do components of the Yamamoto (wheat germ and yeast) diet interact to give the diet varying degrees of gel firmness. This type of approach allows us to understand not just WHAT works in our diets, but more importantly HOW or WHY it works (or fails to work).

In this class and in all Professor Cohen’s classes, we strive to learn the mechanisms, rationale, and basic science behind our rearing systems’ components. We strive to learn HOW and WHY things work as they do in terms of the biology of the insects in our systems, the physical and chemical nature of the interactions of rearing systems’ components, and in knowing the HOWs and WHYs, we are better able to engineer our systems and generate standard operating procedures (SOPs) that are data-based and well-vetted.

If this approach sounds interesting to you, please see the next Blog Page for details on course contents and registration details for the May 2026 class.