The three numbers a sea urchin roe report must keep separate
Sea urchin gonad index is a mass fraction, not an automatic measure of roe quality or saleable yield. A report that says an urchin has a gonad index of 15% means that its measured gonads represented 15% of the defined whole-animal mass at the time and under the weighing protocol used. It does not by itself show that 15% of an incoming commercial lot will become accepted product. Some gonad tissue can be damaged, rejected for colour or texture, lost during cleaning, or measured with a different moisture state. A good translation preserves those distinctions.
This guide explains how to read sea urchin gonad index, roe colour grades, handling losses, grading percentages and batch marketable yield. It is written for readers interpreting an aquaculture trial, seafood purchasing sheet, scientific results table, or technical report in another language. Every calculation below labels its denominator, mass basis and assumed conditions. The worked records are invented teaching examples, not measurements from a farm or a recommendation to harvest wild animals.
Quick reading route: identify the species and sampling date; find the complete wet body-mass definition; locate the gonad weights and their individual denominators; calculate gonad index; separate visual colour and texture grades from weight; reconcile gross roe against cleaned and marketable roe; account for rejected material and missing mass; and state what the evidence cannot prove about taste, food safety or future harvest performance. Those checks are more valuable than carrying a single impressive percentage into a translation without context.
Start with the biological object, not the spreadsheet
Sea urchins are echinoderms, and the edible product often called uni is principally gonad tissue. It is not a shellfish muscle and does not behave like the peeled meat yield of a shrimp or the fillet yield of a fish. An urchin’s hard test, spines, gut contents and coelomic fluid contribute to an incoming whole-animal weight but not to cleaned roe. The proportions change with species, size, feeding, reproductive condition and handling, so mass comparisons must remain attached to a specific sampling protocol.
A technical report may use gonad, roe, uni and reproductive organ in slightly different ways. For a literal translation of a methods section, write ‘gonad tissue’ when the organ is weighed, and retain ‘marketable roe’ when a sorting or packing step has already excluded unsuitable portions. Translating both simply as ‘roe weight’ can erase the very processing loss the author set out to measure.
Some edible lobes are attractive orange or yellow; other tissue may be pale, brownish, granular, watery or otherwise unsuitable for a buyer’s grade. None of those appearances alone gives a bacterial count, toxin result, or legal suitability judgment. Product grading and food-safety release are independent questions. A grade sheet ought to name its criterion and its assessor rather than let the reader invent a market standard.
The right denominator for gonad index
The most common operational gonad index is 100 multiplied by gonad wet weight divided by whole-animal wet weight. Yet ‘whole-animal’ is not self-explanatory. Was the animal weighed alive after drainage? Was attached water allowed to run off? Were any spines trimmed? Were weights recorded before or after a holding period? A translated table should retain the sampling protocol or explicitly label it unavailable, because percentages from different weighing procedures are not necessarily comparable.
Imagine an animal weighed at 320 grams under a defined drained-whole-weight protocol. Its five collected gonad lobes weigh 46.4 grams together. The gonad index is 46.4 divided by 320, multiplied by 100, or 14.5%. A correct unit label is ‘14.5% of drained whole-animal wet mass’. It is not 14.5 grams, it is not 14.5% of the shell-free tissues, and it is not yet a packing yield. The arithmetic checks because 0.145 times 320 grams equals 46.4 grams.
The mass balance is meaningful only for the same individual or same matched group. If average gonad weight comes from twelve animals but average total weight comes from twenty-four different animals, dividing those averages gives a descriptive ratio of unmatched observations, not an individual mean gonad index. Say how specimens were paired. Translators should resist silently ‘cleaning up’ a confused method into a statistically precise claim.
Mean of individual indices is not ratio of pooled weights
Suppose two animals weigh 200 and 500 grams, while their gonads weigh 20 and 50 grams. Each individual index is 10%, and the simple mean of individual indices is 10%. Here the pooled result also happens to equal 70 divided by 700, or 10%. That coincidence does not make the two calculations identical by definition. For unequal individual proportions, the pooling method weights larger animals more heavily.
Try a second pair: animal A is 200 grams with 30 grams of gonad, so its index is 15%; animal B is 500 grams with 40 grams, giving 8%. The arithmetic mean of animal indices is (15+8)/2, or 11.5%. The pooled index is 70/700 times 100, or 10%. Both figures can be correct for different questions. The first describes the average animal in this two-animal sample; the second describes the portion of combined wet biomass occupied by gonads.
In an English translation, ‘mean GI’ is not enough. Prefer ‘arithmetic mean of individual gonad indices’ or ‘pooled gonad mass as a percentage of pooled whole-animal mass’. The phrase may look longer, but it prevents a procurement analyst from treating the biological average as the material yield of an entire shipment.
Wet mass, dry mass and hydration
An urchin report may provide a dry-matter percentage for feeds or a moisture concentration for gonad tissue. Neither belongs automatically in a wet-weight gonad index formula. If a sample of gonad weighs 50 grams wet and contains 20% dry matter by mass, its estimated dry mass is 10 grams. Dividing 10 grams by a 300-gram wet animal produces a mixed-basis value of 3.33%, which is not directly comparable with a 16.67% wet-mass gonad index. The dry matter may be useful for nutritional composition, but the base of comparison has changed.
Draining, rinsing, holding and blotting can change measured wet mass without representing equivalent changes in dry tissue. A buyer examining two lots should ask whether both lots used the same post-harvest preparation. Never translate a ‘wet drained weight’ column as ‘net content’ if packaging water and cleaning steps have not been resolved. Words like rinsed, drained, blotted, shucked, trimmed and packed identify different stages of the mass ledger.
Moisture variation also complicates interpreting apparently rapid gain in roe weight. An increase in measured mass over a short holding interval may combine tissue growth, changes in internal water, changing gut and whole-body mass, and sampling noise. A causal claim of nutritional improvement requires more than a before-and-after percentage from a few unmatched animals.
A clear mass ledger for one imaginary harvest
Consider a teaching lot of 120 urchins with a measured drained whole-animal mass of 36.0 kilograms. The gross extracted gonad mass is 5.04 kilograms. Gross roe as a percentage of incoming wet mass is therefore 5.04/36.0 times 100, or 14.0%. The production team then removes 0.35 kilograms of fragments, damaged material and visible nonproduct tissue, leaving 4.69 kilograms of cleaned roe. The cleaning yield relative to gross roe is 4.69/5.04 times 100, or 93.06%.
After commercial grade sorting, 3.95 kilograms meets the contract’s accepted grade; 0.74 kilograms is clean but rejected for that particular grade. The marketable yield relative to the incoming whole lot is 3.95/36.0 times 100, or 10.97%. Marketable share of cleaned roe is 3.95/4.69, or 84.22%. These two percentages answer different questions and should never be printed under the same unlabeled ‘yield’ heading. All quantities here are hypothetical to demonstrate bookkeeping.
The ledger closes at the processing stage because 5.04 equals 0.35 plus 0.74 plus 3.95 kilograms. It does not close the whole-animal mass, because non-gonad tissues and remaining animal structures are outside that particular ledger. A translator should name the accounting boundary before claiming ‘100% recovered’. The untracked non-gonad mass is not a mysterious processing loss if the report never set out to measure it.
Colour and firmness: observations rather than chemical results
Visual colour grades help buyers and researchers describe roe appearance, but the same number can mean very different things across grading systems. In one study, grade 1 might describe bright yellow-orange, smooth and firm roe, while grade 4 describes unacceptable pale, white or brown tissue with a soft granular appearance. That ordered rubric belongs to that study or contract, not to every sea urchin fishery in the world.
The translator’s job is to carry the exact categories, reference photographs or cards, viewing conditions and assessor procedure forward. A description such as ‘bright orange’ is not a substitute for an instrumental colour coordinate, and an instrumental L*a*b* reading is not interchangeable with an assessor’s opinion of market acceptability. When colour charts differ, retain the original scale and explain the mapping; do not quietly change ‘grade B’ from one buyer into ‘class 2’ from another.
Texture can change with reproductive stage, feeding and post-harvest conditions. The word ‘firm’ might describe how lobes hold their shape under gentle handling; ‘granular’ describes visible or tactile breakdown. Neither proves unsafe food. Likewise a high gonad index can coexist with poor texture. Translate the two results as separate dimensions, not as one composite ‘quality percentage’ unless the original publication defined exactly how to calculate such a score.
A subtle denominator error in grade percentages
A table states that 80% of graded roe is grade A. That sentence may hide at least three denominators: percentage of individual lobes assigned to A, percentage of whole animals yielding any A, or percentage of cleaned roe mass that meets A. Those figures are rarely equal. If small pieces dominate the count but large pieces dominate kilograms, piece-count and weight shares tell different stories.
Imagine that a batch has 120 accepted A pieces weighing 1.8 kilograms and 80 B pieces weighing 2.2 kilograms. By piece count, A is 120/200 or 60%. By mass, A is 1.8/4.0 or 45%. A report that provides ‘60% grade A’ without its denominator cannot support a 60% revenue assumption if sales price is charged by the kilogram.
Readers translating a product sheet should ask whether rejected roe was included in the denominator. If 4.0 kilograms are graded A and B, and another 1.0 kilogram was rejected before grading, the A fraction of all initially clean material is 1.8/5.0 or 36%, not 45%. The exclusion step matters even when the simple fraction is arithmetically flawless.
Sampling design: which animals entered the calculation?
A commercial report can be distorted before a calculator is opened. If large healthy animals are deliberately selected for dissection and small animals remain unmeasured, the sample may overstate expected lot yield. An honest translation names the selection rule, the number of animals examined, and any size, location or holding-tank groups. A total of fifty weighed gonads cannot be assumed to represent five thousand delivered urchins without a defensible sampling design.
A stratified sample can help when animals arrive in two distinct size classes, but a combined estimate still requires the proportions of the larger population in each class. An analyst should not average the two sample means equally if 80% of the incoming lot belongs to one class and 20% to the other. For the mass-yield question, the denominator should follow the actual mass distribution, not simply the number of strata.
Under a translation workflow, add a note such as ‘sample of 30 hand-selected large animals; representativeness not established’ rather than writing ’30 representative samples’ on the strength of guesswork. That single cautious sentence may prevent an expensive misunderstanding. Do not invent a confidence interval if the sampling and dispersion data are missing.
Interpret uncertainty without false precision
The formula for a percentage is simple, but its inputs have error. If a scale records whole mass to the nearest gram and gonad mass to the nearest tenth of a gram, that rounding uncertainty propagates into the reported index. Instrument calibration, drainage time and damage in dissection may matter more than the last decimal place. A table may report 14.527%, but that display does not prove accuracy to one-thousandth of a percent.
For a worked illustration, suppose the gross gonad mass is 48.0 grams with a plausible measurement uncertainty of 0.5 grams and the animal mass is 320 grams with an uncertainty of 2 grams. At the central values the index is 15.0%. If we test the bounding combination of low gonad/high body mass, we get 47.5/322 times 100, or about 14.75%. The high gonad/low body combination gives 48.5/318 times 100, or 15.25%. These are scenario bounds, not a formal statistical confidence interval.
Qualify the distinction in an article: instrument resolution is not the same as biological variability; a measurement interval is not the same as a confidence interval for the population mean. Translate ‘standard deviation’ and ‘standard error’ accurately, and include n, the number of independent observations, where possible. Never turn a standard error bar into a claimed range of all possible animals.
Feeding trials and the temptation to declare a winning diet
Researchers have tested both kelp and formulated diets to improve the marketability of roe from purple sea urchins. A NOAA-hosted research paper describes gonad-index increases under specific experimental feeding conditions, including an index above 15% for selected treatments. That is useful evidence of possibility in the studied species and system, not a promise that every farm will reach the same threshold after the same number of weeks.
Trial descriptions must identify baseline condition, species, tank, ration, feed dry matter, trial duration and sampling stage. If diet A reaches a mean index of 16% and diet B reaches 14%, it may be tempting to translate the difference as ‘two percent more yield’. The absolute difference is two percentage points; the relative index difference is (16−14)/14, approximately 14.3%. Neither alone tells us how many kilograms of grade A uni were sold, or whether differences are statistically credible.
Stocking, tank environment, season and handling may confound simple comparisons. If all treatment animals share one tank, the experimental unit may be that tank rather than the number of animals. A sentence about ‘significantly better feed’ cannot be justified solely because many individual urchins were sampled from one unreplicated tank. Preserve limitations that affect the conclusion.
How to describe quality without pretending it is universal
An aquaculture study may set a marketable threshold tailored to its buyer. Another buyer may reject the same colour, especially when different species, product presentations or destination markets are involved. Say ‘meets the study’s grade-A definition’ or ‘accepted under this purchase contract’ rather than announcing an objective universal A grade. The local standard and sampling conditions are part of the evidence.
The exact expression ‘gonad yield’ can mean the wet extracted organ mass fraction in a biological paper but the packed saleable product fraction on a commercial ledger. This is a translation trap because the English words sound nearly identical. A useful methods box should define gross gonad yield, cleaned roe recovery, grade-A yield, and net shipment yield independently and give the appropriate denominator.
A report is strongest when a reader can reverse the numbers. If the author claims 12% marketable yield from 40 kilograms of whole animals, 4.8 kilograms of accepted product must exist in the ledger. If accepted product is 3.6 kilograms, either the denominator, percentage, or reported category changed. The task is to resolve that mismatch, not polish it away.
Worked example: a buyer’s two-supplier comparison
Supplier North quotes 500 kilograms of delivered wet urchins with 70 kilograms of gross roe and 55 kilograms of grade-accepted roe. Supplier South quotes 300 kilograms of delivered wet urchins with 48 kilograms of gross roe and 36 kilograms of grade-accepted roe. North has a gross gonad mass fraction of 14%, and a marketable whole-lot yield of 11%. South has a gross fraction of 16% and a marketable yield of 12%. South therefore delivers a higher marketable fraction, although North supplies more total kilograms of accepted roe.
Now suppose supplier North charges 2 currency units per kilogram of incoming whole animals and South charges 2.4. North’s raw material cost per accepted kilogram is (500×2)/55, or about 18.18 currency units. South’s is (300×2.4)/36, or 20 units. Under these invented assumptions, the supplier with the better yield has the higher raw-cost per accepted kilogram. Labour, logistics, quality, food-safety controls and selling price have not yet been included.
A correct translation reports both mass performance and commercial cost on their own bases. It does not state ‘South is more economical’ merely because 12% exceeds 11%. Any purchasing recommendation also depends on consistent grade definitions, receiving procedures, live condition, and wastage after the quoted processing boundary.
Worked example: recovery rate is not gonad growth
A trial reports an initial gross roe mass of 4 kilograms from a lot and a later gross roe mass of 6 kilograms from another lot under the same holding system. The difference is 2 kilograms, a 50% increase relative to the first gross roe mass. That phrase alone does not establish individual growth. The number and body mass of animals might have differed, or the later sample might have been larger or healthier.
If each lot contained precisely 100 animals and the mean drained whole mass had been recorded, the reader could calculate per-animal gonad mass and individual indices. Even then the comparison would be of two cohorts unless the same individuals were followed by a validated non-destructive method. Destructive gonad weighing normally precludes tracking the organ of the same urchin at two harvests.
Translate ‘between-batch difference’ when that is what the methods show. Reserve ‘growth’ for a defensible biological comparison, with assumptions plainly stated. This protects readers against attractive but unsupported narratives about feed conversion, ripening or rehabilitation.
Worked example: mass units and decimal places
Imagine a laboratory log showing body mass as 0.275 kg and gross gonad weight as 39.6 g. Converting the body weight to 275 g yields a gonad index of 39.6/275×100, or 14.4%. Dividing 39.6 by 0.275 without the unit conversion produces a nonsensical 14,400% result. A simple bilingual table should put both masses in the same unit before the formula and show the calculation at least once.
If the record uses a decimal comma, 39,6 g represents 39.6 grams in many locales; it is not 396 grams. Likewise 1.250 can mean one thousand two hundred fifty in one presentation and 1.25 in another. Preserve the source number and locale in the audit trail, convert using a declared numerical convention, and have a second reviewer verify high-value weights.
Show rounding explicitly. At 14.4% it may be honest to display one decimal place, especially when weights came from a scale with limited resolution. Another document that reports 14.400000% has not gained six digits of physical truth. Translating measurement precision is part of translating meaning.
Quality assurance before publication or a buying decision
A reliable editorial check starts with the title, then moves through the methodology, tables and conclusions. Are all percentages labelled? Are body masses and extracted gonads measured on compatible wet bases? Are grades and rejection categories defined? Are animal counts, batch weights and statistical replicates recorded? Do the reported equations reproduce the printed results? These questions should be answered before the reader gets to promotional language about premium quality.
Cross-check each total against the listed parts. Ensure that cleaned roe plus cleaning discard equals gross extracted roe, and that marketable roe plus grade reject equals cleaned roe, allowing for known reweighing or moisture adjustments. If the totals do not reconcile, flag a discrepancy with its absolute mass, rather than claiming fraud or inserting an unexplained loss category.
One can translate a technically confused report faithfully while clearly identifying the confusion. Add an editorial note: ‘The original document does not identify whether the 15% yield uses gross or accepted roe.’ That is more useful than pretending perfect certainty. Maintain source terms in brackets when a technical distinction cannot be resolved, and ask the data owner to clarify before contract use.
Transfer problem one: six grades in a mixed box
A receiving clerk has a carton containing 12.0 kilograms of cleaned roe. Contract grade A weighs 4.5 kilograms; grade B, 3.0; grade C, 2.0; rejected colour, 1.2; rejected texture, 0.8; and unclassified samples reserved for laboratory checking, 0.5. First confirm the ledger: 4.5+3.0+2.0+1.2+0.8+0.5 equals 12.0 kilograms. If the question asks for grades A and B as a percentage of all cleaned roe, the answer is 7.5/12.0×100, or 62.5%. If the laboratory-held sample is excluded from the graded denominator, the answer becomes 7.5/11.5×100, or approximately 65.22%.
Neither answer can replace the other without a statement about the laboratory-held material. The issue is the defined population, not arithmetic difficulty. If rejected colour and rejected texture overlap rather than being mutually exclusive categories, the ledger would also double count; the clerk must know whether each piece was assigned to exactly one reason. A robust translation of ‘accepted fraction’ names the classification rule, numerator and denominator.
Transfer problem two: a seemingly excellent gonad index
A pilot group reports a 19% gonad index from twenty selected animals. The incoming commercial run averages 13% when measured across several unsorted bins. Does the pilot prove that the farm can consistently sell 19 kilograms of uni for every 100 kilograms of urchins? No. The pilot describes a selected biological sample’s organ proportion. The commercial run’s measured index describes a different population, and neither figure accounts for removal, cleaning, colour grading, damage or shipping rejection.
The claim can be repaired with a chain of observed measurements: defined whole-animal incoming mass, gross extracted gonad mass, cleaned product mass, accepted product by grade, and finally shipped net content. Each step should have a time, unit and measurement procedure. When a percentage is labelled ‘potential yield’, it should remain hypothetical until the downstream stages are measured. Conditional language is not weakness; it is the exact boundary of the available data.
Transfer problem three: a claimed 30% improvement
A paper states that gonad index increased from 10% to 13%. Calculate the absolute change: 13−10 equals 3 percentage points. Calculate the relative change: (13−10)/10×100 equals 30%. An abstract may correctly say ‘30% relative increase in gonad index’, but a headline that says ’30 percentage points’ is wrong. It would imply an index of 40% if the baseline were 10%. Conversely the phrase ‘3% increase’ is ambiguous because it could suggest either an index rising to 10.3% or a three-percentage-point rise.
Now test a separate claim: marketable kilograms went from 6.0 to 7.2 for equal incoming lot masses. That is a 20% relative increase in accepted product, not the 30% increase in gonad index. Both changes might arise in the same trial without contradicting one another. Do not smooth the difference into one summary number. Record which attribute changed, on what scale, and under which conditions.
Transfer problem four: price per kilogram of accepted roe
A hypothetical processor buys 200 kilograms of whole urchins at 3 currency units per kilogram. Processing costs 140 units, packaging costs 80, and refrigeration and delivery cost 60. It sells 21 kilograms of roe accepted under a defined grade. Gross purchase cost is 600 units; total listed cost is 880 units. Cost per accepted kilogram under those included costs is 880/21, approximately 41.90 units. If the expected selling price is 45 units per kilogram, the stated contribution before unlisted overhead and taxes is (45×21)−880, or 65 units.
If only 18 kilograms remain accepted on arrival, and no listed costs change, cost per accepted kilogram rises to 880/18, about 48.89 units. The apparent gross sale contribution becomes 45×18−880, or minus 70 units. The difference is driven by post-processing accepted mass, not necessarily the biological gonad index. Use ‘under these assumed costs and selling prices’ when translating the result. Never imply that such simplified teaching figures predict a real market’s costs or prices.
Transfer problem five: pooling seasons
An enterprise sampled spring and autumn production separately. Spring delivered 20 kilograms of graded roe from 200 kilograms of whole urchins, or a 10% marketable yield. Autumn delivered 42 kilograms from 300 kilograms, or 14%. The simple average of the seasonal percentages is 12%. The combined yield is (20+42)/(200+300)×100, or 12.4%. The weighted result is suitable for describing the combined incoming mass stream, while the simple average treats the two seasons as equally weighted observational units.
Suppose the report announces 12% as ‘annual marketable yield’ without disclosing that the seasons supplied different masses. This may be an approximation or a wrongly interpreted mean. Translate the sampling scheme and calculate both values for the reader; do not rewrite the underlying data to force a single one. In a longer series of shipments, weight by actual incoming masses when the question is whole-biomass conversion.
Transfer problem six: an incomplete rejection ledger
A packhouse says it extracted 8.4 kilograms of gross gonad tissue, cleaned 7.9 kilograms, and accepted 6.2 kilograms. It reports 1.0 kilogram of colour rejects and 0.4 kilogram of texture rejects. How much cleaned roe remains unexplained? Cleaned minus accepted equals 1.7 kilograms. Listed grade rejects total 1.4 kilograms. Thus 0.3 kilograms remains unassigned after the cleaning stage. The arithmetic is 7.9−6.2−1.0−0.4=0.3 kilograms.
Possible explanations include unlabelled samples, incidental wastage, later reweighing or a recording error; none can be asserted as the true cause. A translated quality report should state ‘0.3 kg not reconciled by the categories provided’ and request clarification. Labelling this automatically as evaporation would be invented. Reconciliation works only when every category covers a mutually exclusive and exhaustive set, or when overlaps and measurement differences are separately explained.
Transfer problem seven: a table copied into another language
A source table labels three columns ‘whole wet weight (g)’, ‘gonad wet weight (g)’ and ‘GSI (%)’. A translation drops the word ‘wet’, and its footnote moves below an unrelated table. The reader can now accidentally combine dry tissue composition with wet animal weight. This is not cosmetic. Restore the complete headings, place the method note adjacent to the table, and retain the original abbreviation in a definition line if the source uses GSI for gonadosomatic index.
If a study used gonadosomatic index in a strict reproductive-biology sense but another study used gonad index for a commercial roe mass fraction, explain their formula rather than force both into one term. Abbreviations travel badly across disciplines. Check all units, multiplication factors, sample numbers, standard-deviation symbols, grade keys, table cross-references and date formats in the translated document. A good editorial review tests the formulas using at least two source rows, not simply their typography.
Frequently asked questions and defensible answers
Is a higher index always better? No. It records a greater measured gonad proportion under a given wet-weight method; buyer acceptance can depend on colour, texture, taste, species, contamination controls and handling. Can colour grade be converted into a bacterial safety score? No. Visual grade and microbiological assurance answer different questions. Can a single tank’s outcome be generalized to an entire coastline? No. External validity depends on species, site, season and replicated design.
Does a study’s 15% marketable threshold apply everywhere? No. Such cutoffs are study- or market-specific and need their scope stated. Why might an index fall even if gonad mass is steady? Whole-animal mass could rise, changing the denominator. Why can the pooled yield differ from average animal index? Heavier animals contribute more to the pooled mass ratio. Is the term ‘uni’ always precisely equivalent to freshly dissected gonad? Product usage varies; state whether the document concerns raw organ, cleaned roe, graded product or packed product.
What should the translator do when a variable is undefined? Retain the original term, label the denominator as unavailable and ask for the methods or purchase specification. What if reported totals do not close? Present a numerical discrepancy, preserve original values, and request a correction rather than manufacturing an explanation. What if the text contains a legal or food-safety claim? Verify against the relevant jurisdiction’s actual rules and testing requirements; a biological yield guide cannot certify compliance.
A practical teaching checklist that follows the material
Begin with one imaginary animal and verify body mass, gonad mass and index by direct substitution. Next bring together two unequal animals to see why arithmetic mean of individual indices and pooled index answer different questions. Expand to a batch ledger in kilograms, adding gross extraction, cleaning, grading and net shipped mass. Each extension changes the denominator, so write it explicitly next to every percentage. The calculation must remain understandable even if a reader sees only one screenshot of the table.
Next test the language. Ask a second reader to describe what ‘yield’, ‘recovery’, ‘grade A’ and ‘marketable’ mean without seeing the translator’s notes. If their answers are inconsistent, the definitions need strengthening. Finally check the source: not every study, supplier and region shares the same species, biological condition, colour grading scale or acceptable endpoint. Preserve the methods’ real constraints and avoid universalising a result from one trial.
The finished report should let an independent reader recompute each important number, distinguish measured from estimated outcomes, identify missing evidence and reach a cautious conclusion. That is the practical standard for translating technical sea urchin roe records: neither literal terminology alone nor polished sales language is enough. The translation has to protect the scientific and commercial meaning of the measurements.
Primary sources and related reading
NOAA repository: purple sea urchin gonad enhancement study
University of Western Australia: algae supplements and gonad grades