Nutrient Adequacy and Biological-Architecture Classification: A Two-Axis Framework for Evaluating Variable Fresh-Food Diets in Companion Animals

The BARC Protocol · Release B-2.1.0 · September 2026

Author. Prasanna M, Growlrr Foods Pvt Ltd, Chennai, India. Correspondence: support@growlrr.com. ORCID: 0009-0003-6812-347X.

DOI. Cite 10.5281/zenodo.22648597 — the concept DOI, which always resolves to the current version of this framework. Use it whenever you mean BARC itself. This particular version is 10.5281/zenodo.22648598; cite the version DOI only when you need to point at exactly the text you read.

Relationship to the series. This is the framework. The eight manuscripts deposited alongside it are one implementation of it, and are not the framework; they are listed at the end.

Published. 8 September 2026

Licence. Creative Commons Attribution–ShareAlike 4.0 International (CC BY-SA 4.0) — free to use, adapt, implement and commercialise, with attribution, and adaptations of this document carry the same licence. Practising the protocol requires no licence at all: a method and a set of thresholds are not copyrightable, so nothing here restricts formulating against it, manufacturing to it, or selling what you make. The licence governs reuse of this text. Full terms accompany the record.

Terms. A glossary of the statistical and nutrition vocabulary used across the accompanying series — requirement, floor, safe upper limit, Wilson bound, N-limited, draw, stress multiplier, retention factor, trituration and the rest — is in the series cover note (10.5281/zenodo.22647606), section 4.1.

Computational tooling and AI assistance. The model is implemented in TypeScript under Node.js with no external numerical dependencies. Claude (Anthropic) was used as a coding and analysis assistant in developing this work and preparing these records; it was not used to generate results, all of which were produced by executing the model. See the series cover note (10.5281/zenodo.22647606), section 6.1.

Plain-language material. A non-technical statement of the protocol is at growlrr.com/barc-protocol; the scientific framing at growlrr.com/barc-protocol-scientific-framework; the case for correcting a home-prepared bowl at growlrr.com/why-balance-homemade-pet-food; and what "complete and balanced" means against a published standard — the adequacy axis of this protocol — at growlrr.com/pet-nutrition/complete-and-balanced. The public reference list behind these claims is at growlrr.com/pet-nutrition-references.

Status. A framework standard published by Growlrr Foods Pvt Ltd and shared openly for citation and scrutiny. Not peer-reviewed as of the date of publication, 8 September 2026, and not a consensus position of any external body.

Competing interests and funding. The author is the founder of Growlrr Foods Pvt Ltd, a pioneer and proponent of the BARC Protocol for pet foods. BARC stands for Biologically Appropriate, Residually Corrected, and is conceptualised as an open framework under a Creative Commons licence, for like-minded collaborators around the world to formulate and develop. Growlrr Foods Pvt Ltd develops and commercialises a BARC-based residual-correction product, BowlBalancer. The work was funded entirely by the company; no external funding was received. No independent laboratory, institution or reviewer has verified the results reported here. Readers should weigh these facts alongside the analysis, which is published in full — including its limitations and open items — precisely so that it can be checked.

Abstract

Standard regulatory compliance models evaluate companion-animal diets with a single scalar test: does each nutrient clear a defined floor? That test is indifferent to the source of the nutrient and to the long-term appropriateness of the food supplying it. We argue that the quality of a companion-animal diet resolves along two independent axes — biological appropriateness and nutritional adequacy — and that conflating them is the central confusion in the fresh-feeding debate.

We define BARC (Biologically Appropriate, Residually Corrected) as a diet meeting both requirements: a fresh-food base clearing published quantitative boundaries for appropriateness, and a calculated residual correction closing the measured gap to an established nutrient target (NRC 2006, and by extension FEDIAF and AAFCO profiles). Neither alone qualifies. We give the boundaries numerically, derive the four diet states the two axes produce, and demonstrate with a deterministic point-estimate model why a one-axis check is insufficient: a meat-free, grain-based bowl can read "adequate" while its total sulfur amino acids sit at roughly 111% of the recommended allowance and collapse to 61% when a single animal-protein component is removed — a structural failure no micronutrient correction can repair.

We also set out the quantitative reason the two axes come apart in a modern household. Requirements for mass-tied nutrients track body mass; intake tracks calories. A free-ranging carnivore running a multiple of a sedentary pet's maintenance energy receives that multiple of every micronutrient against a requirement of the same order, so an ancestral recipe is adequate at ancestral energy throughput and incomplete at domestic throughput. We show that the organ-meat lever conventionally offered to close that gap is mathematically closed: in a cat bowl the vitamin A ceiling is reached at a liver fraction of roughly 16%, while the manganese floor is not reached until roughly 47%.

Keywords: BARC Protocol · homemade dog food · homemade cat food · companion animal nutrition · fresh feeding · home-prepared diet · nutritional adequacy · biological appropriateness · diet classification · residual correction · sulfur amino acids · NRC 2006 · energy throughput

1. Introduction — the fresh-food adequacy gap

The modern fresh-feeding movement rests on a sound intuition: cats are obligate carnivores, dogs are facultative carnivores, and both evolved eating fresh animal tissue rather than shelf-stable extruded kibble. Diets built around fresh muscle, organ and egg have measurable appeal — palatability, digestibility, and owner confidence in a short, recognisable ingredient list.

That intuition is not a formulation. A diet can look ancestral and still be incomplete. The foundational evidence is Stockman et al. [1], who evaluated 200 home-prepared maintenance recipes for dogs drawn from 34 sources, 129 of them written by veterinarians. Five recipes met the NRC recommended allowances for all essential nutrients and nine met the AAFCO profiles — so 195 of 200 (97.5%) failed at least one NRC allowance. The companion feline study analysed 94 recipes and found that none met the NRC recommended allowances for every essential nutrient, most falling short on three or more and some on as many as nineteen [21]. Stockman et al. also tested the usual defence — that rotating varied recipes averages out to completeness — and found that it did not [1]. The gaps cluster where a cook cannot see them — calcium, iodine, zinc, copper, the B-vitamins. Around that spine sits a clinical literature of predictable consequences: taurine-deficiency cardiomyopathy and retinal degeneration in cats [2,3], hypervitaminosis A from liver-heavy feeding [4], pansteatitis from unbalanced fish-based diets [5], and the broader risk profile of raw meat-based diets [6].

The regulatory answer is the language of "complete and balanced" — a food meeting the AAFCO nutrient profiles or an equivalent FEDIAF or NRC target [7,8,9]. This is a genuine guarantee, but it is a guarantee about adequacy, not appropriateness. A diet can be certified complete while deriving most of its energy from plant starch and rendered fractions.

The field therefore talks past itself. Advocates of fresh feeding emphasise appropriateness and under-deliver on adequacy; advocates of formulated feeding emphasise adequacy and dismiss appropriateness as marketing. Both describe real, separable properties. This paper separates them explicitly and defines a standard requiring both.

2. Why the two axes come apart in a modern household

There is a quantitative reason, and it rests on an asymmetry between two things usually conflated.

The composition of prey is fixed per kilogram; its micronutrient density does not change with the eater. Requirements are set per day. Adequacy of an uncorrected whole-prey diet is therefore a function of energy throughput: a free-ranging carnivore running a multiple of a sedentary pet's maintenance energy eats that multiple of the same food and runs a surplus of most micronutrients against a requirement of the same order. Whether that surplus stays under ceilings depends on organ fraction; §2.2 shows that it does not for vitamin A above roughly 16% liver. The whole-prey ratio is not "nature's balance"; it is nature's balance at nature's calorie burn. The domestic animal — neutered, indoor, ageing — lives on a small energy base and runs at the margin, and the low-density micronutrients fall through.

The proposition this framework rests on is that domestication changed the energy throughput rather than the requirement — the table above is the arithmetic that motivates it.

This is not the claim that needs scale with energy — for most mass-tied nutrients they do not. The B-vitamins partly do ride energy flux, which is why the per-kilocalorie convention suits them. The thesis is strongest on the mass-tied set, and the mass-tied set is exactly the set that fails in the arithmetic below.

2.1 Demonstration A — adequacy is a property of throughput, not of the recipe

Take a generic declared whole-prey raw recipe of the kind widely sold — approximately 92% muscle and connective tissue, 7% organ, 1% bone — evaluated brand-agnostically from public composition data [10] against this framework's floors, with floors held fixed per day and only intake scaled:

At a threefold throughput the recipe clears every micronutrient floor it fails at pet throughput. That multiple sits above the receipted range (§2.3), so the reading has to be the narrower one: the shortfall tracks energy throughput rather than the recipe, and at realistic wild multipliers it narrows without closing. Vitamin D sits near zero at any multiple on standard composition panels — the wild animal's vitamin D arrives via prey fat and liver stores and sun, neither of which scales indoors; that row is noted, not leaned on.

2.2 Demonstration B — the organ lever is mathematically closed

"Add organs" is the standard answer. Parametrise the liver fraction λ of a cat bowl (baseline 10%, liver displacing muscle one for one, 251 kcal/day, this framework's floors, and the FEDIAF vitamin A maximum of 100,000 IU per 1,000 kcal [8], itself derived from the feline hypervitaminosis-A case series of Seawright et al. [4]):

The ceiling arrives 2.8 times before the floor. In this parametrisation — liver displacing muscle one for one, at 251 kcal/day, against this framework's floors — there is no liver fraction at which the bowl is simultaneously manganese-adequate and vitamin-A-safe, and copper breaches on a parallel path because liver is the densest copper source in the bowl. This is why a "complete" claim on an uncorrected whole-prey product routes through a premix, disclosed or not.

2.3 The wild multiplier, receipted — and smaller than intuition

The sustained metabolic scope of free-ranging animals is reported as mostly between 1.5 and 5 times resting [11]. A modelled feral cat at roughly 300 kcal/day [12] against a neutered indoor cat at 52–75 kcal·kg^−0.67 [8] gives about 1.6–2.3×, not 3×; sedentary versus working dogs (70–95 versus 150–175 kcal·kg^−0.75) about 1.6×.

At the receipted 1.6–2.3× the manganese row of §2.1 reaches only 51–74% of floor. Volume narrows the gap; it does not close it. Diversity of prey is the other half of the wild mechanism, which is why the recommended base is a rotation (§5.1) and why a single frozen grind is one draw from nature's menu rather than the menu.

Reference 12 is a modelled nutrient profile, not a doubly-labelled-water measurement, and is described as such throughout. No doubly-labelled-water study of feral cats was found.

2.4 The standards bodies already print the thesis

FEDIAF's adult-dog values are the NRC 2006 requirements "increased by 20% to compensate for the lower energy requirement of household dogs" compared with the intake NRC assumed [8, §3.3.1]. Its guidelines carry a dual-column structure in which the lower-energy animal receives the higher per-1,000-kcal floor for every nutrient of concern:

The stated reason is the thesis in the standard's own words: "the energy needs may be satisfied before the requirements of protein, minerals or vitamins are met… Hence a systematic adjustment applied to all essential nutrients is needed when fed below the NRC standard assumption" [8, §7.2.5]. The corollary the standard concedes is the whole argument: any fixed-composition food, prey included, that satisfies the high-energy column can fail the low-energy column with no change in the food.

FEDIAF's 20% covers the household-versus-NRC-assumed gap, not household-versus-feral. It establishes the direction and the mechanism, not the full multiple.

3. The BARC definition

BARC stands for Biologically Appropriate, Residually Corrected. A diet is BARC if and only if it is:

1. a biologically appropriate fresh-food diet — meeting the quantitative boundaries of §4; and
2. residually corrected — a calculated correction has been added to close the measured gap between the fresh base and the daily nutrient target.

Both conditions are required.

Provenance — BARC stands on two named precedents.

The principle of biological appropriateness comes from BARF — Biologically Appropriate Raw Food — coined by the Australian veterinarian Dr Ian Billinghurst, whose Give Your Dog a Bone (1993) [13] brought evolutionary, prey-model feeding to a mass audience. BARF supplied the biological axis and the insight that a diet's resemblance to what the species evolved on is a real property worth defending.

The discipline of correcting a home diet to a published nutrient standard comes from BalanceIT, the home-formulation system created by Dr Sean Delaney, DVM, MS [14,15], a board-certified veterinary nutritionist and former UC Davis faculty member. BalanceIT demonstrated that a home-prepared diet can be completed to an NRC or AAFCO target by calculation rather than guessed at — that the arithmetic is tractable, and that an owner will follow it if it is made simple enough.

BARC is the synthesis, and it adds five things.

1. Both axes, as independent and mandatory requirements. BARF holds appropriateness and is largely silent on measured adequacy; a correction-only approach holds adequacy and is indifferent to what the food is made of. BARC requires both and publishes quantitative boundaries for the first, so that appropriateness stops being an adjective and becomes a test a diet passes or fails.
2. A stochastic validation framework rather than a point estimate. A recipe that meets the requirement on paper, at nominal ingredient values, has been shown to meet it once. BARC asks what fraction of realistically varying daily bowls clears every floor, and reports the answer as a confidence bound with its draw count and basis stated.
3. Unit dosing. The correction is a whole sealed sachet, dosed by weight class, not a scoop or a spoon (§5.4). This removes the domestic measurement error that a microgram-scale instruction otherwise guarantees.
4. Customisable diets within a fixed structure. The base is a diverse animal rotation the owner can substitute within — white meat, red meat, fish, egg — with the correction unchanged (§5.1, §5.2), and the framework re-implementable against a different regional food supply entirely.
5. A kitchen-counter experience that a person will actually sustain. The daily instruction reduces to a count of sachets for an animal of a given weight. Everything else — the residual computation, the allometry, the twenty-odd nutrient floors — happens before the owner is involved, which is the only place that complexity can survive contact with a real household.

One deliberate departure from BARF: BARC is boneless and cooked. Calcium and trace minerals are delivered by the correction rather than by feeding bone, removing the choking, gut-perforation and raw-pathogen risks that bone-in raw feeding carries, and removing the per-meal variance of an ingredient whose mineral content is not knowable.

4. The two axes, the four tiers, and the boundaries

• Axis 1 — Biological appropriateness. Does the diet match the metabolism a carnivore evolved for? A property of the food base only; the correction plays no part in it.
• Axis 2 — Nutritional adequacy. Does the complete bowl meet an established requirement standard at the animal's life stage? A property of the whole bowl including the correction.

Each tier claims only what its own criteria test. The correction's incidental palatant protein is counted in the adequacy panel but excluded from the Axis 1 test.

"Complete and balanced" is not the ceiling of diet quality; it is a welfare floor any responsible diet must clear. Biological appropriateness is a second, higher requirement layered on top.

4.1 The boundaries

In this framework appropriateness is a set of quantitative boundaries a candidate diet either clears or does not. They are not derived from NRC, AAFCO or FEDIAF adequacy profiles, which govern Axis 2; they are anchored to the empirical literature on ancestral prey composition and on what these species select and tolerate.

No adequacy standard supplies these numbers, and none can. The NRC, FEDIAF and AAFCO profiles specify nutrients, not sources: crude protein, methionine plus cystine, arachidonic acid, taurine, calcium, and so on down the list [7,8,9]. Nothing in any of them requires that a nutrient arrive in animal tissue. A plant- or cereal-based diet with its deficits made good by purified amino acids, a fatty-acid supplement and a vitamin-mineral premix — sprayed or blended onto the base — can satisfy every line of a profile and be labelled complete and balanced, correctly.

This is not a loophole being exploited; it is the standards working exactly as designed. They are deliberately source-agnostic so that they can govern any manufacturing route, and that neutrality is what makes them useful as a floor. But it has a consequence that is rarely stated: an adequacy standard has no vocabulary in which to say that a diet is appropriate for an obligate carnivore. It cannot express the claim, because it does not describe sources at all. A diet can therefore be fully compliant and still bear little resemblance to anything the species evolved to eat — and conversely, an ancestral bowl can be manifestly appropriate and fail the profile badly, which is what §1 documents.

The same omission applies to relationships between nutrients, not only to sources. The profiles set a minimum for crude protein and a minimum for fat, and in places a maximum; they specify a small number of ratios, most notably calcium to phosphorus. They do not specify a protein-to-fat relationship. A diet can therefore clear the protein minimum and the fat minimum simultaneously while sitting anywhere across a very wide range of macronutrient balance, and remain fully compliant throughout. This is not hypothetical: free-choosing domestic dogs converge on an intake near 30% of energy from protein and 63% from fat — a protein-to-fat ratio near 0.48 [19] — which is compliant, palatable, self-selected, and for a sedentary urban animal obesogenic. A standard built from independent per-nutrient floors has no means of objecting to it, because no floor has been breached.

Sources and relationships are the two things an adequacy profile deliberately leaves open. They are also, between them, most of what "biologically appropriate" means. That gap is the whole reason this framework carries a second axis. If appropriateness is to be assessed at all, someone has to state numerically where its boundaries lie, because no standards body does. BARC states them, and states them as its own.

These five thresholds are proposed, not measured, and the distinction is load-bearing. Each is positioned by reference to a cited anchor, but the anchors bound the values rather than determine them: nothing in the literature fixes the canine animal-energy floor at 85% rather than 82% or 88%. They are the framework's stipulations — a deliberate, arguable choice of where to draw a line — and they are stated numerically so that they can be disputed with evidence rather than debated in the abstract. Anyone implementing BARC against a different food supply, or with better data than was available here, is expected to challenge these numbers, and the framework is versioned so that they can move.

Animal-derived energy. The floors are anchored on ancestral prey composition: the free-ranging wolf derives close to 99% of dietary energy from animal sources [16,17]. The BARC floors are set deliberately below that anchor to admit a realistic mixed bowl while remaining unambiguously carnivore-appropriate; the exact offset is a judgement. The asymmetry reflects biology: the cat is an obligate carnivore with metabolic dependencies — pre-formed taurine, arachidonic acid, vitamin A, limited carbohydrate handling [7]; the dog is facultative with bounded capacity to use plant fractions.

Protein-to-fat ratio. A single envelope for both species. The ancestral wolf diet supplies approximately 52% of energy from protein and 47% from fat, a ratio near 1.1 [16]; adult cats offered free choice converge on roughly 52% protein and 36% fat, a ratio near 1.44 [18]. Both sit inside one 0.8–1.6 window, so the species difference does not live on this axis. The envelope is bounded on both sides: below 0.8 the diet drifts fat-forward toward an obesogenic balance; above 1.6 it usually signals a lean mono-protein construction rather than a fat-bearing rotation.

Appropriate is not the same as preferred. Free-choosing domestic dogs converge on a fat-forward intake near 30% protein and 63% fat by energy, a ratio near 0.48 [19] — below the appropriateness floor and, for a sedentary urban dog, obesogenic. The boundary is anchored on the biologically appropriate ratio, not the modern preferred one. "Let the animal choose" is not a formulation strategy.

The basis must always be stated. The validated reference bowls run 0.85 (dog) and 0.82 (cat) by energy; the same diets expressed by mass are near 1.9, because protein carries 4 kcal/g and fat 9 kcal/g.

Fibre. Whole prey carries very little fibre. The dog ceiling corresponds to the upper end of the 2–5% dry-matter crude-fibre range described for small-animal maintenance [20,8]; the cat ceiling is tighter because the obligate carnivore evolved on an almost fibre-free diet. This is a line the framework chooses: AAFCO requires crude fibre to be declared but sets no maximum [9], so a rice-and-legume bowl at roughly 10% dry-matter fibre can be AAFCO-compliant and inappropriate at once.

5. Residual correction, in concept

5.1 The fresh base — diversity as a requirement, not a preference

BARC prescribes a diverse base rather than a single protein: red meat, white meat, fish and egg, together, as part of the daily diet. In the reference implementation that is mutton or goat; chicken; farmed white fish; egg; and, for dogs, curd and a low-oxalate vegetable mix fed for satiety and fibre rather than for micronutrients.

Diversity is a requirement of the framework because no single protein supplies a complete micronutrient spectrum. Rotating across complementary proteins narrows the residual gap before any correction is applied, and it spreads the risk of any one ingredient being atypical. A single-protein bowl leans harder on the correction and keeps a narrower margin; it is a welfare floor, not the standard.

Fat is an essential part of the diet, not a residue of the protein. It carries the fat-soluble vitamins and the essential fatty acids, and it is the reason the protein-to-fat envelope is bounded on both sides (§4.1) rather than treated as "more protein is better". A very lean bowl is a failure of the base, not a virtue of it.

5.2 Substitution — the framework travels

The named ingredients are an implementation, not the standard. BARC does not prohibit substitution within a class, provided the energy is corrected:

• White meat — chicken may be exchanged for turkey or duck
• Red meat — mutton or goat for beef or pork
• Fish — tilapia for cod or another white fish

What the framework fixes is the structure — a diverse animal base clearing the appropriateness boundaries, with the residual gap closed by calculated correction. What it leaves open is which specific animals fill that structure.

The reference implementation names what is commonly and affordably accessible across India: a country with long coastlines, large landlocked interiors, and established farmed fresh-meat supply. Those choices are a response to that supply, not a claim that those species are uniquely correct.

BARC is intended to be re-implemented regionally. A formulator elsewhere should populate the base from what is accessible, affordable and culturally acceptable in their own market, recompute the residual against the same nutrient standard, and arrive at a different ingredient list satisfying the same framework. That portability is the point of publishing the framework openly rather than as a product specification: the arithmetic transfers even when the shopping does not.

5.3 Decoupling — the owner may feed to satiety

The single most consequential property of residual correction is not chemical. It is that the dose is set by the animal's body weight, not by how much food is served.

The unit count is computed from an allometric function of body weight alone. There is no energy term for cats; for dogs there is only a floor-defence term that increases the count at low metabolic rate. Serving more food, or less, does not change how many units the animal receives.

Three things follow, and together they are the practical case for the whole framework.

The animal can eat to appetite. Because macronutrients and micronutrients are supplied by two independent channels, the owner can increase or reduce the protein and fat in the bowl — for appetite, for body condition, for a working day versus a rest day — without changing the micronutrient delivery at all. In a fixed-composition food the two are welded together: every adjustment to the ration is simultaneously an unintended adjustment to twenty-odd nutrients.

A heavy animal can be put on a diet safely. This is where the welding does real harm. Restricting a fixed-composition food to reduce weight restricts every micronutrient in proportion, at exactly the moment the animal can least afford it — which is why weight-loss diets are conventionally reformulated to be nutrient-dense rather than simply fed in smaller amounts. Under residual correction the reduction happens only on the grocery side; the correction is unchanged, so micronutrient density per calorie rises as the ration falls. Weight management stops being a nutritional trade-off.

It removes the daily micro-measurement problem. The alternative, for an owner genuinely trying to complete a home-prepared diet, is to measure and combine on the order of two dozen separate micronutrients every day — the requirement panel used here runs to twenty-three, and the model floors twenty-eight — in milligram and microgram quantities, from separate products with their own overlaps and gaps. That is not realistic at a kitchen counter, and attempting it is how both under-supply and accidental over-supply occur. A single weighed unit replaces the entire exercise, and it also replaces the ad-hoc supplement stack that usually accumulates around a home-prepared diet.

The bound. This holds across the validated envelope — the metabolic multiplier sampled over [1.2, 1.8] for dogs and [0.9, 1.3] for cats — and the reliability figures are computed with that variation included, not at a single point. It is not a licence to feed arbitrarily: the base still has to clear the appropriateness boundaries of §3, and the correction closes a residual computed against a specified base, not against any bowl whatsoever.

5.4 The unit-dose format

The correction is delivered as whole, single-use sachets — a fixed composition, a declared mass, a whole number per day, one size per weight band, with no splitting, no scooping and no mixing of sizes.

The format is doing real work. A scoop or a spoon reintroduces exactly the measurement variance the framework exists to remove, and at microgram scale a part-scoop is not a measurable quantity in a domestic kitchen. A whole sealed unit also protects the labile actives until the moment of use, and it makes the delivered dose auditable after the fact: what the animal received is a count, which an owner can report accurately and a formulator can reason about, rather than an estimate of a partial measure.

It also makes the instruction trivial: the owner's entire micronutrient task is "this many sachets a day for an animal of this weight."

But the claim that matters is not that the instruction is easy to follow — it is that the formulation does not depend on the instruction being followed precisely. A real kitchen is not a laboratory. Portions are eyeballed, one protein is swapped for another because of what was in the shop, the fish is oilier this week than last, a meal is early or late or skipped, and the cooking is hotter or longer than the recipe assumed. A correction that only holds when the household behaves consistently is not a solution to a household problem.

The framework's answer is to require that robustness be demonstrated rather than asserted, by deliberately breaking the assumption in the model. In the validation of the accompanying implementation, the day-to-day variability of the biological inputs, the grocery composition and the cooking retention is inflated by a stress multiplier and the whole population is re-run. The multiplier is applied specifically to the kitchen coefficient of variation — the day-to-day wobble in what actually goes in the bowl. Two results come out of it. On a settled daily pattern the reliability of the corrected diet does not move at all: it holds at 99.996% from realistic variability all the way to fourfold it, for both species. On the harder alternate-day protein-rotation pattern it does degrade — to 99.391% (cat) and 99.421% (dog) at fourfold — but it degrades gracefully, and it still clears the 99% release gate at four times the variability any real kitchen is likely to produce (Manuscript 2, Appendix A).

Both halves of that are worth stating. The correction is not indifferent to how erratic the household is; it is simply built with enough margin that ordinary inconsistency, and a good deal more than ordinary inconsistency, does not push the diet below its floors.

That is the property being claimed, and it is the reason the correction is a fixed unit rather than a scoop. The daily micronutrient delivery is insulated from the daily rhythms of the kitchen — not because owners are expected to be precise, but because the design does not spend its margin on their precision.

Why the precision belongs in a facility. The framework corrects with pharmaceutical-grade actives rather than with botanicals or organ tissue, and §2.2 and Manuscript 4 §4 give the reason: a purified active has a declared assay, while kelp, liver and similar sources vary by species, season, site and lot in ways that are invisible to the formulator because they are not declared per lot. Choosing the controlled input is the right call, but it carries an obligation that the framework states plainly.

Several of these actives are toxic when handled imprecisely. Cholecalciferol, selenium compounds and iodine sources have a narrow interval between the daily requirement and a harmful dose, and they are dosed in micrograms — quantities no domestic scale resolves, and which a household cannot triturate uniformly through a carrier by eye. A few milligrams of error, undetectable on a kitchen counter, separates a correct dose from an injurious one. The variance argument and the safety argument therefore point the same way: the accuracy has to live in a high-precision blending facility, and what reaches the household should be a count rather than a measurement.

This is why BARC is offered as an open framework for formulators and manufacturers, and not as a home recipe, and why implementations of it — including the one specified in the accompanying series — should publish the delivered nutrient specification, which is what makes an adequacy claim checkable, while not publishing a percent-by-weight preparation that would read as instructions for compounding these materials at home. Anyone implementing BARC is asked to hold the same line. The framework is open so that it can be built on, examined and argued with; it is not an invitation to weigh microgram actives in a kitchen.

5.5 The residual

For a given animal and base, the framework computes the daily nutrient target from an allometric model [Manuscript 4] against the NRC 2006 requirement set [7], estimates what the base delivers, and defines the residual as the shortfall, nutrient by nutrient:

residual(nutrient) = daily target(nutrient) − fresh-base contribution(nutrient), where positive.

Residual correction is the act of dosing that shortfall rather than a whole nutritional profile. It does not mean that every nutrient lands at 100% of its requirement, and it cannot: the unit is a fixed composition dosed by count, so the nutrient with the largest relative shortfall sets the unit count, and nutrients with smaller shortfalls or wide safe margins are consequently delivered well above their floors — long-chain omega-3 reaches roughly 760% of requirement in the worked bowl of §6, against a nutrient with no established ceiling at these intakes. What the principle constrains is what is corrected and how it is dosed; the operative limit on the overshoot is the safe upper limit rather than the floor, which is why the margin to every ceiling is computed explicitly [Manuscript 6]. The residual is dosed allometrically: the count of identical units scales with the animal, never the dose within a unit.

What the correction does not contain is as informative as what it does. It carries no protein-building amino acids — no methionine, no cystine, no lysine, no tryptophan — and no arachidonic or linoleic acid. This is not an omission; a correctly specified animal-protein base supplies the amino-acid and omega-6 requirements without assistance, and the paired-arm ablation bears this out: the uncorrected bases fail on minerals and vitamins, not on protein or amino-acid adequacy (Manuscript 8 §4). A diet that has to have its amino acids sprayed back on is one whose base was not appropriate to begin with. That is the practical difference between residual correction and reformulating a profile from a cheap base: get the base right and the macronutrients come free, so the correction can address only what the base cannot supply.

Two nutrients in the correction sit outside the strict micronutrient class, and both are there for a measured reason rather than a categorical one. Long-chain omega-3 (EPA and DHA) is present because the uncorrected bowls deliver 0% of the requirement — a fresh bowl built on farmed white fish and land animals does not carry it (Manuscript 8 §4). Taurine is present because it is a genuine feline shortfall: on the modelled bowls the food alone supplies about 63% of the feline requirement, against roughly 250% of the canine, which is why the feline correction carries several times the canine taurine load. The rule the framework follows is empirical, not categorical: correct what the appropriate base measurably fails to deliver, and nothing else.

5.6 Interaction hazards that a per-nutrient checklist does not catch

Most published accounts of home-diet failure are reported nutrient by nutrient: this diet was short of calcium, that one of zinc. A second class of failure behaves differently. It is created not by omitting a nutrient but by a decision made elsewhere in the bowl, and it is invisible to a checklist that evaluates each nutrient independently. Three cases recur often enough in fish-containing and raw home diets to be worth stating explicitly, because in each the owner is usually doing something they believe to be beneficial.

Vitamin E, and the requirement that moves when omega-3 is added. The vitamin E requirement is not a fixed number: polyunsaturated fatty acids consume vitamin E through oxidation, so the requirement rises with polyunsaturated intake. A bowl can therefore be vitamin-E-adequate before fish oil is added and deficient after — and the owner adding the oil is doing so deliberately, for the omega-3 benefit, with no indication that they have just moved a different requirement. The clinical endpoint is pansteatitis, the painful inflammation of adipose tissue historically called yellow fat disease, which is well described in cats on unbalanced home-prepared and fish-heavy diets [5]. Vitamin E is, in our view, the most commonly overlooked nutrient in exactly the diets that are most careful about omega-3.

Thiaminase, and raw fish. Several fish species — carp and herring prominently — carry thiaminase, an enzyme that destroys thiamine. Cooking denatures it; raw feeding does not. Cats fed raw carp or herring develop clinical thiamine deficiency within weeks, and the presentation is neurological rather than digestive, so it is not readily attributed to the diet by an owner [22]. A raw fish-forward bowl can therefore be formulated to contain adequate thiamine on paper and still deliver a deficiency, because the destruction happens in the bowl rather than in the calculation. This is why the framework's retention model assigns thiamine a retention factor of zero — it counts no dietary thiamine at all, from any source, and requires the correction to supply the entire requirement on its own. That choice was made for cooking loss, but it happens to be the correct defence against thiaminase too: a nutrient you never counted cannot be destroyed out from under the calculation.

Vitamin K, and fish-predominant diets. This one is not a subtlety at all — it is an explicit requirement most home feeders have never encountered. The AAFCO cat food nutrient profiles require a source of vitamin K in any diet containing more than 25% fish on a dry-matter basis [9]. A household building a fish-forward bowl from whole ingredients has no reason to know that a threshold exists, still less where it sits.

What these three have in common is that each is a consequence of an otherwise sensible choice — add fish oil, feed raw, favour fish — and none is detectable by checking nutrients one at a time against a table. They are arguments for correcting a diet as a system rather than as a list.

The framework's answer is to move all three into the correction, at a fixed load, and out of the household's hands. The correction described in the accompanying series carries vitamin E, thiamine and vitamin K₁ as standing components, sized against the specified base rather than added when someone remembers. What reaches the kitchen is a sachet that is torn and poured once a day; the retention modelling, the interaction accounting and the manufacturing control sit behind that, where they can be done properly and audited.

One boundary should be stated with it, because it is a real limit of any fixed-composition correction. The vitamin E load is sized against the specified diet, including that diet's own omega-3 content. A household adding substantial further oil beyond the specified base moves the polyunsaturated load, and therefore the vitamin E requirement, outside what the fixed correction was sized for. The decoupling described in §5.3 applies to feeding more or less of the specified bowl; it does not extend to adding arbitrary quantities of a nutrient-active ingredient on top of it.

5.7 A validation floor is not a diet

For robustness testing the framework uses a deliberately austere single-protein (chicken-only) construction as a worst-case floor: it stresses the correction harder than any realistic rotation. Chicken-only is a stress case, not a recommended diet, and no claim in this series endorses single-protein feeding.

6. Results — the technical demonstration, and why one axis is insufficient

Using a deterministic point-estimate model of the correction engine — an illustration, not the Monte-Carlo validation gate of Manuscripts 3 and 8 — take a meat-free, grain-based home bowl of white rice, lentil dal, curd and mixed vegetables, with one egg, and add the correction.

Under a one-axis adequacy check the bowl reads "in envelope." The correction fills the micronutrient gaps; rice and dal fill the energy target. Essential nutrients run from roughly 120% to 760% of their floors, and long-chain omega-3 reaches 762% of requirement with no fish in the bowl, because it is carried entirely by the correction. The correction architecture works — and that is precisely why correction alone is not a sufficient test.

Under the two-axis check the same bowl fails Axis 1 — animal-derived energy far below the boundary, fibre roughly double the ceiling — and is classified Adequate — not BARC. It also reveals a structural fragility the floor test hides. Because the framework scores protein by amino-acid quality rather than crude protein, methionine plus cystine (total sulfur amino acids) is the single binding constraint of any grain-and-legume bowl: cereals are sulfur-adequate and lysine-poor, pulses are lysine-rich and sulfur-poor, and the pair surviving that complementation as limiting is methionine plus cystine.

With the egg present, total sulfur amino acids sit at about 111% of the recommended allowance — a knife edge, with crude protein at the floor. Remove the egg and they fall to 61% of the recommended allowance and 76% of the NRC minimum: sub-base, a failure the food must fix, because a micronutrient correction carries no bulk sulfur amino acids. What "passes" a one-axis test is an ovo-lacto diet on a rice filler, balanced on the two components that happen to be animal products.

For cats the wall is different and harder. Taurine and arachidonic acid are animal-tissue nutrients; cats cannot elongate linoleic acid to arachidonic acid; and the correction's omega-3 is marine EPA and DHA, which does not supply it. A plant-protein or egg-white-only cat bowl is therefore sub-base regardless of how much correction is added.

Between the floor and the safe upper limit, adequacy is binary. A nutrient at 380% of requirement is not "better" than one at 110%; the excess is only margin.

7. Founding principles

The taxonomy is the framework's structure; the four principles below are positions taken within it, stated as authored convictions grounded in the disclosed mechanisms, not as external findings.

1. Diversity over the mono-diet. A varied animal-protein rotation spreads nutritional risk; a single-protein diet leans harder on correction and offers a narrower margin. A welfare floor, not a recommendation.
2. Nutrition is not a cliff. Carnivore intake is physiologically buffered on both ends. On the lower end what matters is intake averaged over rolling physiological windows — B-vitamins over roughly 7–10 days, most minerals over roughly 30, copper and iron over roughly 90–180 [Manuscript 6]. On the upper end chronic accumulation is held under safe upper limits by hepatic storage and excretion. Missing one day does not break adequacy; one rich day does not breach the ceiling.
3. Healthy animals self-regulate. A healthy adult, given real activity, regulates its own energy balance. The framework sizes the ration to real needs, leaves headroom, and treats body condition reassessed over weeks as the governing signal.
4. Activity is the norm. Movement is the baseline a carnivore's metabolism assumes; the sedentary urban animal is the deviation. Energy is sized to real activity and condition [Manuscript 4] — a retriever on a thirty-minute walk is not a wolf — and restoring activity, not cutting the bowl, is the first lever for an overweight animal.

8. Scope and limitations

• Healthy animals only. This is a framework for healthy adult and growing cats and dogs. It is not a clinical or therapeutic system, makes no disease claims, and is not a substitute for a veterinary therapeutic diet. Animals with diagnosed disease should be fed under veterinary direction.
• The boundaries are choices, not derivations. No standards body sets an animal-derived energy floor or a fibre maximum for companion animals. This framework holds a fibre ceiling where AAFCO holds none, and says so.
• The taxonomy is an authored classification, and the principles of §7 are convictions grounded in disclosed mechanisms, not external findings.
• The demonstrations of §2 rest on public composition data [10] and this framework's floors, and describe the category — declared meat, organ and bone — not any named product. No brand-level comparison is made or should be inferred.
• Wild adequacy at high throughput means "clears these floors", not "wild diets are optimal." Wild carnivores are not a welfare benchmark.
• Nothing here claims fresh feeding is better than commercial feeding. The argument is scoped to completeness, never to freshness. There is no consensus that switching to a raw or fresh diet improves outcomes, and this paper does not assert one.
• §6 is a deterministic point-estimate illustration, not the validation gate. The stochastic validation is Manuscripts 2 and 8.
• Not peer-reviewed as of the date of publication (8 September 2026). Should it later be reviewed or independently replicated, that will be recorded in a subsequent version of this record.

9. An open invitation — how to implement, test and challenge BARC

BARC is published as a framework to be built on, not as a specification of a product. It is licensed CC BY-SA 4.0: anyone may implement it, adapt it, teach from it and publish against it, commercially or otherwise, provided attribution is given and adaptations of this document are released under the same terms. The ShareAlike condition exists for one reason — to stop the written protocol being enclosed by any implementation of it, including ours. A commons that can be taken private is not a commons.

It is worth being exact about who that condition binds, because ShareAlike is often over-read. Practising BARC requires no licence and no permission from anyone. A method, a taxonomy and a set of thresholds are not copyrightable; nothing here restricts formulating against this protocol, manufacturing to it, or selling what you make, and no obligation to open a formulation arises from doing so. The licence governs reuse of this document's text — so it binds someone who republishes or adapts the specification itself, and nobody else.

The most useful thing we can offer a collaborator is a specific list of what is unresolved. Each item below is a real gap, and work on any of them would improve the framework rather than merely endorse it.

1. Implement it against a different food supply. The parameterisation here is built on what is reliably available in India — a coastline, a large landlocked interior, farmed poultry and freshwater fish. The framework's structure should be independent of that, but the claim is untested. A Mediterranean, Scandinavian, Andean or West African implementation, with locally available proteins and locally realistic energy budgets, would establish whether the two-axis structure travels or whether it silently encodes Indian ingredient availability. We would consider a failure here a genuinely valuable result.

2. Challenge the stipulated boundaries with evidence. The five appropriateness thresholds in §4.1 are proposed, not derived, and we have said so. They are stated numerically precisely so that they can be argued with. Evidence that would move them: intake or outcome data placing the animal-derived energy floor elsewhere; a defensible protein-to-fat envelope from feeding rather than self-selection studies; a fibre ceiling grounded in gastrointestinal outcome rather than ancestral analogy. A paper arguing that the canine floor belongs at 78% rather than 85%, with data, is exactly the response this document is written to invite.

3. Test the results independently. No external laboratory, institution or reviewer has verified anything reported in the accompanying series. The validation methods are specified in enough detail to be reimplemented from scratch, and reimplementation by someone with no stake in the outcome would be worth more than any further internal run. Where our numbers are wrong, we would rather know.

4. Close the empirical gap. The framework's quantitative claims are modelled. They rest on published composition tables, published requirements and stated assumptions, and they have not been measured in animals. Non-invasive household observation, veterinary practice data, or a controlled feeding study would all bear on this, and none of it exists yet. This is the framework's largest weakness and we would rather it were addressed by someone else than by us alone.

5. Extend or contest the taxonomy. The two-axis claim — that appropriateness and adequacy are independent, and that conflating them is the central confusion in the field — is the framework's core assertion, and it is the thing most worth attacking. A demonstration that the axes are not independent, or that a single-axis treatment suffices, would refute the framework rather than refine it. That is a legitimate and welcome outcome.

6. Species beyond cats and dogs. Nothing in the reasoning is intrinsically feline or canine; the residual-correction logic should apply to any species whose ancestral intake assumed an energy throughput its domestic form no longer has. Whether it does is an open question.

How to engage. Cite the concept DOI 10.5281/zenodo.22648597 when you mean the framework. Work that implements, extends, contradicts or supersedes it is welcome to say so plainly, and we will link independent implementations and published critiques from the framework's public page. Correspondence to support@growlrr.com.

A word on disagreement. This document is authored by a commercially interested party, states that openly, and will be read with that in mind — as it should be. The appropriate remedy for that conflict is not for us to be more emphatic; it is for the framework to be checkable, checked, and corrected where it is wrong. Published disagreement is therefore a contribution to BARC, not an objection to it.

10. Conclusion

The argument of this protocol is that the quality of a companion-animal diet resolves along two axes that are routinely treated as one. Nutritional adequacy asks whether each nutrient clears a floor. Biological appropriateness asks what the food is. Adequacy standards are deliberately indifferent to the second: they specify nutrients rather than sources, and levels rather than relationships between them. A cereal base with purified amino acids and a vitamin-mineral premix sprayed on can satisfy every line of a profile, correctly. No standards body publishes an animal-derived energy floor or a protein-to-fat envelope, because it is not their job to.

That gap is why this protocol carries a second axis and states its boundaries numerically. Those five thresholds are stipulated, not derived — positioned against cited anchors that bound them without determining them — and they are given as numbers precisely so that they can be disputed with evidence rather than debated in the abstract.

The practical claim that follows is narrower than it is often taken to be. An appropriate fresh base supplies the macronutrients without assistance; what it does not reliably supply is a set of micronutrients, and correcting only that residual is a smaller and more tractable problem than formulating a complete diet. Whether the boundaries proposed here are the right ones is an open question, and §9 sets out what evidence would move them. BARC is not owned by any implementation of it, including the one specified in the accompanying series.

References

1. Stockman J, Fascetti AJ, Kass PH, Larsen JA. Evaluation of recipes of home-prepared maintenance diets for dogs. J Am Vet Med Assoc. 2013;242(11):1500–1505. doi:10.2460/javma.242.11.1500
2. Pion PD, Kittleson MD, Rogers QR, Morris JG. Myocardial failure in cats associated with low plasma taurine: a reversible cardiomyopathy. Science. 1987;237(4816):764–768. doi:10.1126/science.3616607
3. Hayes KC, Carey RE, Schmidt SY. Retinal degeneration associated with taurine deficiency in the cat. Science. 1975;188(4191):949–951. doi:10.1126/science.1138364
4. Seawright AA, English PB, Gartner RJW. Hypervitaminosis A and deforming cervical spondylosis of the cat. J Comp Pathol. 1967;77(1):29–39. doi:10.1016/S0021-9975(67)80004-5
5. Niza MMRE, Vilela CL, Ferreira LMA. Feline pansteatitis revisited: hazards of unbalanced home-made diets. J Feline Med Surg. 2003;5(5):271–277. doi:10.1016/S1098-612X(03)00051-2
6. Freeman LM, Chandler ML, Hamper BA, Weeth LP. Current knowledge about the risks and benefits of raw meat-based diets for dogs and cats. J Am Vet Med Assoc. 2013;243(11):1549–1558. doi:10.2460/javma.243.11.1549
7. National Research Council (US), Committee on Animal Nutrition. Nutrient Requirements of Dogs and Cats. Washington (DC): The National Academies Press; 2006. doi:10.17226/10668
8. FEDIAF. Nutritional Guidelines for Complete and Complementary Pet Food for Cats and Dogs. Brussels: FEDIAF; 2024. https://europeanpetfood.org/self-regulation/nutritional-guidelines/ (This is the edition on which the nutrient tables used here are based; a 2025 edition was published on 8 September 2025 and is not the basis of these runs.)
9. Association of American Feed Control Officials. Dog and cat food nutrient profiles. In: AAFCO Official Publication. Champaign (IL): AAFCO; 2016. (The 2016 profiles are the edition used by the model.)
10. Haytowitz DB, Ahuja JKC, Wu X, Somanchi M, Nickle M, Nguyen QA, et al. USDA National Nutrient Database for Standard Reference, Legacy Release [dataset]. USDA ARS; 2019. doi:10.15482/USDA.ADC/1529216
11. Peterson CC, Nagy KA, Diamond J. Sustained metabolic scope. Proc Natl Acad Sci U S A. 1990;87(6):2324–2328. doi:10.1073/pnas.87.6.2324
12. Plantinga EA, Bosch G, Hendriks WH. Estimation of the dietary nutrient profile of free-roaming feral cats: possible implications for nutrition of domestic cats. Br J Nutr. 2011;106(Suppl 1):S35–S48. doi:10.1017/S0007114511002285
13. Billinghurst I. Give Your Dog a Bone. Bathurst (NSW): Warrigal Publishing; 1993. ISBN 0-646-16028-1
14. Fascetti AJ, Delaney SJ, editors. Applied Veterinary Clinical Nutrition. 1st ed. Ames (IA): Wiley-Blackwell; 2012. doi:10.1002/9781118785669. (A second edition was published in 2024; the first edition is the one consulted here.)
15. Delaney SJ. Balance IT [online home-diet formulation tool]. Davis (CA): DVM Consulting; 2003–. https://balanceit.com (accessed 7 September 2026).
16. Bosch G, Hagen-Plantinga EA, Hendriks WH. Dietary nutrient profiles of wild wolves: insights for optimal dog nutrition? Br J Nutr. 2015;113(S1):S40–S54. doi:10.1017/S0007114514002311
17. Buff PR, Carter RA, Bauer JE, Kersey JH. Natural pet food: a review of natural diets and their impact on canine and feline physiology. J Anim Sci. 2014;92(9):3781–3791. doi:10.2527/jas.2014-7789
18. Hewson-Hughes AK, Hewson-Hughes VL, Miller AT, Hall SR, Simpson SJ, Raubenheimer D. Geometric analysis of macronutrient selection in the adult domestic cat, Felis catus. J Exp Biol. 2011;214(6):1039–1051. doi:10.1242/jeb.049429
19. Hewson-Hughes AK, Hewson-Hughes VL, Colyer A, Miller AT, McGrane SJ, Hall SR, et al. Geometric analysis of macronutrient selection in breeds of the domestic dog, Canis lupus familiaris. Behav Ecol. 2013;24(1):293–304. doi:10.1093/beheco/ars168
20. Merck Veterinary Manual. Nutritional requirements of small animals. Rahway (NJ): Merck & Co. https://www.merckvetmanual.com (accessed 7 September 2026).
21. Wilson SA, Villaverde C, Fascetti AJ, Larsen JA. Evaluation of the nutritional adequacy of recipes for home-prepared maintenance diets for cats. J Am Vet Med Assoc. 2019;254(10):1172–1179. doi:10.2460/javma.254.10.1172
22. Markovich JE, Heinze CR, Freeman LM. Thiamine deficiency in dogs and cats. J Am Vet Med Assoc. 2013;243(5):649–656. doi:10.2460/javma.243.5.649

How to cite

Prasanna M. (2026). Nutrient Adequacy and Biological-Architecture Classification: A Two-Axis Framework for Evaluating Variable Fresh-Food Diets in Companion Animals (The BARC Protocol). Growlrr Foods Pvt Ltd. Zenodo. https://doi.org/10.5281/zenodo.22648597

Status of this document

This protocol is deposited under a Creative Commons Attribution–ShareAlike licence, and deliberately so. BARC is intended as an open framework that others may implement, extend and disagree with. Depositing it gives it a permanent, archived identifier that can be cited in the literature — which a company web page cannot — The concept DOI above always resolves to the current version, so the framework can be revised as collaborators contribute without breaking any citation already made to it. A living, human-readable copy is maintained at growlrr.com/barc-protocol.

Collaboration is invited explicitly. BARC is not owned by any implementation of it, including the one specified in the accompanying series. Formulators, veterinary nutritionists and researchers who wish to implement it in other regions, against other food supplies, or with other correction chemistries are invited to do so and to publish the result. Correspondence: support@growlrr.com.

It is published instead at growlrr.com/barc-protocol under a Creative Commons licence. The eight Zenodo records in the accompanying series are one implementation of this framework — its calcium architecture, energy model, manufacturing method, safety analysis and validation — and each cross-references this framework rather than restating it.