What a fertilisation percentage can and cannot tell you
Sea urchin fertilisation rate is the proportion of examined eggs that show the study’s defined evidence of fertilisation at a stated observation time. It is not automatically the percentage that becomes normal larvae, survives settlement, or reaches a marketable juvenile size. When translating sea urchin spawning records, preserve the original biological endpoint as carefully as the arithmetic. Egg release, egg recovery, sperm exposure, fertilisation, first cleavage and later embryonic normality are successive observations, not interchangeable names for one success rate.
This practical guide helps readers interpret sea urchin broodstock spawning logs, egg-concentration calculations, dilution series, fertilisation microscopy, embryonic staging and batch reconciliation. The worked examples use hypothetical hatchery numbers; they are arithmetic and interpretation exercises, not instructions to conduct biological procedures. Real aquaculture practices depend on species, authorized facilities, water conditions, trained staff and biosecurity requirements. The NOAA Fisheries account of Hawaiʻi’s collector urchin hatchery and official echinoid test methods provide primary evidence for the distinctions discussed here.
Quick reading summary: locate the broodstock species and identity; separate animals stimulated to spawn from animals producing usable gametes; record egg volumes and true cell counts; track each aliquot’s dilution and sampling volume; define a fertilised egg by the protocol and observation time; report abnormal or multiple-sperm events separately; check later cleavage or larval survival against the right starting cohort; and flag missing denominators or uncertain observations before writing any bold success claim.
Broodstock response is an animal count
A spawning sheet may have columns headed ‘animals induced’, ‘responding females’, ‘responding males’ and ‘usable gamete donors’. These need separate denominators. An apparent response rate of six out of twelve adults could mean six of all animals, or six out of eight females, or six out of ten with verified maturity. A translation should never introduce a reproductive sex classification that the observers did not establish. In some urchins external appearance does not allow reliable sex assignment before gamete release.
Suppose a hypothetical run involves 24 animals, of which 18 release visible gametes, while 15 yield collections passing an internal quality check. The initial spawning-response percentage is 18/24×100, or 75%. The usable-donor percentage relative to all animals is 15/24×100, or 62.5%. The usable share among responders is 15/18×100, or approximately 83.33%. None of these numbers gives eggs per female, fertility rate, or family representation.
A hatchery may hold broodstock for more than one day or select only some collections for fertilisation. Do not treat repeated records from the same animal as independent animals. If a donor releases gametes on two occasions, the unit of analysis may be collection event rather than unique individual. Cross-reference donor identity and dates in the technical summary.
Release, collection and usable eggs are not synonyms
The visible release of egg-rich fluid does not measure how many viable eggs were captured. Material may be lost during transfer, excluded by mesh, retained in equipment, or rejected because it contains debris. Source documents may call the collection a ‘spawn’, a ‘suspension’ or an ‘egg stock’. The translator should retain the operational meaning: a volume of liquid with a measured or estimated concentration of eggs at a specified stage.
Imagine a recovered egg suspension with a volume of 2.5 litres and an estimated concentration of 180 eggs per millilitre. Since 2.5 litres equals 2,500 millilitres, the estimated stock count is 2,500×180, or 450,000 eggs. If a later filtration step retains 390,000 eggs by a separately measured estimate, apparent recovery relative to the first estimate is 390,000/450,000×100, or 86.67%. The unaccounted 60,000 are a difference between estimates, not automatically dead eggs.
Volume must be the actual egg suspension volume, not the nominal capacity of the bucket. A collection in a vessel marked ‘5 L’ may contain only three litres. For an honest report write ‘estimated 450,000 eggs from a measured 2.5 L suspension’ instead of ‘450,000 eggs from a 5 L bucket’.
Count chamber arithmetic and the danger of hidden multipliers
A microscope count usually examines a small fraction of a larger suspension. The calculation requires counted cells, the observation volume, any prior dilution and the total stock volume. A number copied from a counting sheet without the chamber dimensions or dilution factor may be impossible to verify. If the observer counts four subvolumes, name whether the sheet reports their sum or average.
Take a hypothetical 0.05 mL sampled chamber containing 40 eggs after a 1-in-4 dilution of the original stock. The observed diluted concentration is 40/0.05, or 800 eggs/mL. The original stock concentration is 800×4, or 3,200 eggs/mL. If the stock volume is 150 mL, the estimated count is 480,000 eggs. Removing the dilution multiplier would understate the estimate fourfold. Treat a ‘1:4’ notation carefully because labs sometimes use it for one part sample plus four parts diluent rather than one part in four total.
State the dilution using actual volumes if possible, such as ‘1 mL original plus 3 mL seawater, final volume 4 mL’. Then the factor is exactly four. A bilingual document that retains ambiguous shorthand without explaining this convention can generate a large and preventable error in stocking density.
Multiple aliquots reveal sampling variability
Counts from different aliquots are not identical. Eggs may settle, clump or move with currents while the sample is taken. If five counts from identical chamber volumes are 48, 51, 39, 56 and 46 eggs, their mean is 240/5, or 48 per chamber. Their spread matters as well; a report that contains only the mean hides variability that may reveal poor mixing or a small counting volume.
Suppose each chamber volume is 0.1 mL with no dilution. A mean count of 48 gives 480 eggs/mL. For 900 mL of stock the estimated total is 432,000 eggs. If an operator instead multiplied the highest single count of 56 by the stock volume, the estimate would be 504,000 eggs, a substantial difference. Neither number should be reported to six meaningful digits without considering counting uncertainty.
Distinguish technical replicate chamber counts from independent biological egg collections. Five aliquots from one donor are useful for estimating one sample’s concentration; they are not five separate female broodstock. Reporting an apparent n=5 biological sample from one suspension is a denominator mistake with consequences for statistical interpretation.
The fertilisation endpoint must be observed, not guessed
Many sea urchin methods examine structures or events associated with fertilisation, such as a fertilisation envelope, rather than waiting for the entire culture to grow. The chosen diagnostic feature, microscope condition and observation time belong in the methods. An official Canadian biological test method uses echinoid fertilisation assays under controlled conditions, where gamete preparation and appropriate controls determine whether the result can be interpreted.
Suppose 200 eggs are scored under a defined microscopy endpoint. The observer calls 172 fertilised and 28 unfertilised. Observed fertilisation proportion is 172/200×100, or 86%. If 15 of those 172 later appear abnormal in a second scoring session, it is not legitimate to subtract 15 from the original fertilisation numerator unless the original endpoint explicitly required normal later development. The second result is an additional conditional outcome.
Translate ‘envelope formation’, ‘cleavage’, ‘normal embryo’ and ‘competent larva’ using distinct terms. A paper that says ‘86% fertilisation’ does not promise that 86% of initial eggs will become healthy juveniles. Keep the endpoint anchored to the measurement.
Sperm concentration is not the number of eggs fertilised
A sperm-stock report may give concentration as sperm per millilitre, a dilution ratio or the volume of diluted suspension added. The number of sperm present is a potential exposure measure, not a count of successful fertilisations. Sperm viability and activation, seawater conditions, time since collection, egg condition and mixing can affect outcomes. Excess sperm exposure can introduce polyspermy risk, so more sperm is not equivalent to better fertility.
A concentration reported as 5 million sperm/mL in a 20 mL stock corresponds arithmetically to an estimated 100 million sperm in that stock. If 0.2 mL is sampled from it, that sample contains approximately one million sperm under the assumption of homogeneous suspension. The calculation does not establish how many eggs those sperm reached or fertilised. Do not translate this total as ‘one million fertilised eggs’.
Ratio language needs care: a 1:100 dilution can mean one part original made up to one hundred parts total or, in ambiguous notes, one plus one hundred. The best report lists measured volumes and final concentrations. Do not convert an official assay’s validated species-specific conditions into universal hatchery operating advice.
Fertilisation, polyspermy and abnormality
A successful sperm-egg interaction is not the only concern when reporting embryos. Multiple sperm entering an egg may disrupt normal development. An early morphological feature may not expose all such failures. A report may therefore keep separate counts for apparent fertilisation and morphologically normal embryos at a later checkpoint. These are nested stages of observation, and their denominator must be stated.
Imagine 1,000 eggs originally distributed to a teaching vessel. At an early check, 900 show the fertilisation marker. At the next independent census, 765 appear normal for the expected stage. One statement is 90% early fertilisation relative to initial eggs. The other is 76.5% normal development relative to initial eggs, or 85% of the early-fertilised estimate if the later count truly followed the same cohort and no major losses or sampling changes occurred.
Do not infer that the remaining 135 all suffered polyspermy; developmental abnormality can have multiple causes, and some eggs may be missing or misclassified. A careful translator keeps cause-of-failure hypotheses separate from observed counts.
Early cleavage follows a developmental clock
An embryo-stage log may give minutes or hours after insemination. Temperature and species affect the developmental timetable, making an isolated statement like ‘normal at 60 minutes’ incomplete. Distinguish clock time, age since sperm exposure, and biological stage. If samples were collected at different temperatures, two records with the same elapsed time may represent different expected stages rather than an error in the microscope.
Suppose 150 observed embryos are checked at an agreed interval and 120 have reached a defined cleavage stage, while 30 remain at earlier stages or are unclassifiable. The stage-attainment percentage among scored embryos is 120/150×100, or 80%. It is not an 80% hatching rate, and it is not a survival rate for the entire initial stock unless the sample is representative and the original denominator is known.
Translations should not substitute ‘hatched’ for ‘cleaved’. Cleavage is early embryonic cell division; hatching from an embryonic envelope, where the species has such a stage, is another event; swimming larval appearance and settlement occur later. These steps can be separated by substantial time and require distinct observations.
An apparent 95% result with a missing sample size
The statement ‘95% fertilised’ looks impressive but cannot be evaluated properly without n. Was it 19 of 20 eggs, or 950 of 1,000? Both are 95% as a point estimate, yet their sampling precision differs. A proportion from a small chamber should not be presented as a certainty about a million-egg tank.
For a simple illustration, a 95% result from 20 scored eggs corresponds to 19 positives. One different classification would change the value by five percentage points. For 1,000 scored eggs, one classification changes the displayed percentage by only 0.1 percentage point. That comparison does not provide a formal confidence interval, but it shows why the underlying count matters.
If the source says ‘three replicate dishes’ with 100 eggs per dish, preserve the dish-level results if possible. Pooling the 300 eggs may hide one poor dish and two good dishes. Labelling the average without explaining whether each dish was weighted equally is another source of ambiguity.
Fertilisation controls and test validity
Official biological assay methods distinguish sample treatments from control seawater and use acceptability criteria before interpreting an effect. This matters particularly when fertilisation is measured to study toxicity, rather than to maximise hatchery production. A high or low outcome in an unknown sample only has meaning relative to a valid control run under the specified method.
Imagine a test treatment produces 70 fertilised eggs among 100 scored, while its control produces 90 among 100. The observed treatment fertilisation is 70%, the control is 90%, and treatment divided by control is 70/90, about 77.78%. The absolute difference is 20 percentage points. Calling the treatment ‘77.78% fertilisation’ is false: that number is a relative-to-control ratio, not the treatment’s raw observed rate.
Never state that a test ‘passes’ a standard by intuition. Verify the exact published method, species, exposure design, controls, acceptability criteria and version. Translators should reproduce the named standard accurately and avoid applying the method’s threshold to unrelated hatchery conditions.
Recovery after washing and what disappeared
In a hatchery, eggs or embryos can be rinsed or screened to remove debris and excess sperm. A net change in estimated count after this operation has several possible explanations: physical loss, retention on mesh, altered suspension volume, counting variability or actual developmental failure. The numbers alone cannot distinguish them. A technical record is stronger when it gives upstream and downstream volumes, aliquot counts and uncertainties.
Consider a hypothetical 500,000 estimated eggs before a transfer and 420,000 estimated eggs after it. The apparent numerical recovery is 84%. That does not mean the other 16% necessarily died. Without observing the missing eggs, writing ‘16% mortality caused by filtration’ would invent mechanism and cause. Report ‘apparent recovery of 84% across the transfer step’ and include the sample uncertainty.
If a count after transfer actually refers only to selected normal-looking embryos, the downstream number has changed both its processing stage and inclusion criterion. The report must distinguish a mechanical recovery calculation from biological selection. The stages of the ledger should each state exactly what is counted.
Family diversity and pooled gamete reporting
Separate donors matter in hatchery genetics. If egg collections from six females are mixed and sperm from four males is pooled, a result from the mixture cannot automatically reveal each donor’s fertilisation success. An individual may have contributed far more eggs or sperm than another. Pooling may be useful for production or a specific research design, but it removes some individual-level information.
Suppose donor A contributes 300,000 eggs and donor B contributes 100,000. If 90% of A’s eggs and 70% of B’s eggs fertilise, the total expected fertilised eggs under this teaching scenario are 270,000+70,000, or 340,000. Overall rate is 340,000/400,000×100, or 85%. The unweighted average of donor rates is 80%. Neither formula should be used without knowing whether egg allocation and fertilisation were measured separately.
A translated phrase such as ‘average fertilisation rate of donors’ is not equivalent to ‘fertilisation of pooled eggs’. Preserve whether the unit is a donor, a family, a dish or a batch. This distinction also protects against overly confident genetic diversity claims from a mere count of participating broodstock.
A single spreadsheet can conceal many timepoints
A row labelled day 1 could contain a collection date, not embryonic age; a date in the tank notebook might use local time while a data export uses UTC. Another file might store minutes after insemination as a decimal hour value. When interpreting a developmental rate, connect each timepoint to the same starting event, and preserve whether a measurement was made before or after a transfer.
Suppose gametes were mixed at 22:50 local time and examined at 00:20 the following day. The elapsed interval is 90 minutes, not negative 22.5 hours and not 24 hours. If the translator strips calendar dates, readers may not see the midnight boundary. The appropriate label is ’90 min post-insemination’ when that is truly the recorded clock interval.
Timing estimates may be approximate if mixing occurred over several minutes or individual dishes were not started simultaneously. A high-quality methods translation tells the reader when the clock began and when each observation was made. Unsupported precision in hours is just another form of mistranslation.
Quality, hygiene and the limits of this document
Eggs and embryos are sensitive biological materials, and official hatchery records may include water temperature, salinity, filtration, bacterial monitoring and handling procedures. Such details help establish comparability and identify possible stressors; they do not support a diagnosis from a spreadsheet alone. The appropriate action for a suspected contamination event depends on site procedures and qualified facility personnel.
A high fertilisation value cannot substitute for disease screening, animal health documentation, environmental permits or genetic management. Likewise a low fertilisation rate does not prove a single culprit. The record might contain information about gamete maturity, timing, temperature, exposure or counting error, but those factors should be weighed with actual observations rather than guessed retrospectively.
Mark the boundary plainly in a translated report: ‘These percentages describe the observed fertilisation endpoint and sampling method; they do not demonstrate subsequent larval survival or suitability for release.’ Such a sentence makes the science easier to use, especially when documents pass between laboratories, hatcheries, conservation programmes and funders.
A complete first-to-last accounting example
An imaginary hatchery records six responding female donors and estimates 1,200,000 eggs in their combined collection. After a documented washing and allocation step, 1,080,000 eggs are estimated in usable stocks. The apparent retention is 1,080,000/1,200,000×100, or 90%. At the fertilisation check, representative microscope samples yield a weighted estimate of 91% early fertilisation. Under the assumption that those samples represent the usable egg stock, the estimated count showing the chosen fertilisation marker is 982,800.
At a later checkpoint, 80% of the early-fertilised estimated cohort reaches a defined normal embryonic stage. That conditional estimate produces 786,240 embryos, which is 72.8% of usable eggs and 65.52% of the first estimated collected eggs. These calculated counts are model-derived estimates, not direct whole-tank enumeration. The arithmetic preserves the stage-specific denominators and makes explicit each sampling assumption.
Some readers would multiply 90% retention by 91% fertilisation by 80% normal development and announce 65.52% ‘survival’. That product is valid only as a chain of conditional proportions under the specified linked-cohort assumptions. The more accurate label is ‘estimated normal embryos as a percentage of initial recovered eggs’, not a lifelong survival or field-recruitment claim.
Transfer exercise: compare two days without changing the population
On Monday a hatchery estimates 280,000 collected eggs and 224,000 of those are retained after washing. On Tuesday it estimates 360,000 collected and 288,000 retained. Both retention fractions are exactly 80%, although Tuesday’s retained count is 64,000 larger. Reporting ‘Tuesday’s recovery improved’ because more eggs were recovered is misleading if the intended metric is percentage retention. The number of eggs recovered increased, but the fraction retained did not. A technical translation should make this distinction obvious.
Now suppose Monday’s retained eggs show 90% fertilisation under the chosen marker and Tuesday’s show 75%. Estimated marker-positive counts are 201,600 and 216,000 respectively. Tuesday has a lower fertilisation fraction yet a higher estimated marker-positive count. The combination is not contradictory. Different initial quantities and percentages can move in opposite directions. Always report both the amount and the denominator-specific performance measure.
Transfer exercise: reconcile a chamber with four count categories
A microscopist scores 250 eggs: 202 with the prescribed fertilisation envelope, 31 without it, 12 damaged beyond confident classification and 5 that could not be observed properly. The scoreable eggs are 202+31=233. Under a protocol permitting exclusion of unscorable material, observed fertilisation among scoreable eggs is 202/233×100, or approximately 86.70%. As a fraction of everything encountered, marker-positive eggs are 202/250×100, or 80.8%. These figures answer different questions.
Do not assume the exclusion is automatically permitted. An official method may prescribe how damaged eggs are handled and which counts invalidate the test. If the test protocol requires the full 250 as a denominator, a translator must not silently switch to 233. In an explanatory report, display both fractions with their definitions and then apply the method’s rule where available. Technical accuracy requires method compliance as well as mathematical consistency.
Transfer exercise: reconstruct the original dilution
A student finds ’50 eggs in 0.02 mL; diluted 1+9′ on a paper form. If one volume of original sample plus nine equal volumes of diluent produces ten total volumes, the counted diluted suspension contains 50/0.02 = 2,500 eggs per mL. The original concentration is then estimated as 25,000 eggs per mL. At a stock volume of 40 mL, the corresponding total is 1,000,000 eggs. The entire estimate depends on the interpretation of the ‘1+9’ notation.
Had the operator meant one volume of original made up to nine total volumes, the factor would instead be nine and the estimate 900,000 eggs. The correct resolution is to check the documented dilution volumes, not choose whichever figure sounds more plausible. Mark the uncertainty in the report if the source record cannot resolve it. Preserve the difference between final volume and added diluent volume when moving between languages.
Transfer exercise: distinguish true sample numbers from technical duplicates
A study displays nine fertilisation percentages: three microscope fields sampled from each of three dishes. A novice translator writes ‘nine independent replicates’. That is not supported. The design contains three dishes, with three subsamples or fields in each, assuming dishes are the experimental units. Averaging nine fields as though all were independent may understate between-dish variation and make a weak treatment effect look more certain.
Calculate first within each dish and report dish-level measures. If dish one fields score 82%, 84%, 86%, its average is 84%. If dish two averages 75% and dish three 90%, the average of dish means is 83% when each dish is weighted equally. The main statistical comparison should reflect the design, not simply the total number of fields. Preserve hierarchical language in methods: animal, donor, pooled collection, vessel, dish and microscope field.
Transfer exercise: detect the wrong basis in normal development
One report claims 80% fertilisation, followed by 75% normal cleavage. The crucial question is whether 75% means of all starting eggs or of the fertilised eggs. With 10,000 starting eggs, the first interpretation gives 7,500 normal-cleavage eggs; the second gives 0.75×0.80×10,000=6,000. Both can appear in the literature, but they are substantially different. The sentence must include the denominator.
Suppose the underlying count sheet lists 8,000 fertilisation-marker eggs and 6,000 normal-cleavage eggs. Then 6,000/8,000=75% of early fertilised eggs, and 6,000/10,000=60% of the original batch. A faithful translation should show the relationship, not compress both into an ambiguous ‘75% normal’. Later development is a conditional outcome, and its percentage should tell the reader exactly what condition was applied.
Transfer exercise: examine a false precision in an estimate
A tank volume is recorded as approximately 190 litres and a concentration estimate as 3.8 larvae per millilitre. Converting litres to millilitres gives 190,000 mL; the estimated total is 722,000 larvae. Writing exactly 722,000 live larvae would exaggerate certainty because the volume is approximate, the concentration is sampled, and survival was not necessarily scored. The honest wording is ‘about 722,000 larvae estimated from the sampled concentration’.
An independent count later finds a different concentration after mixing. The change may reflect biology or heterogeneity in the sampling; one result cannot identify the cause. State whether counts use replicate aliquots, what time they refer to, and how the concentration was derived. A large number of digits in a spreadsheet formula is not a substitute for documented accuracy or validated sampling.
Transfer exercise: identify a failed test without guessing a diagnosis
A hypothetical control dish shows 40 fertilised eggs among 100 examined, while treatment dishes score 32%, 36% and 34%. A comparison of treatment and control numbers alone could lead a rushed reader to label the treatment harmful. But if the official assay requires a higher control fertilisation threshold than this experiment achieved, the test may be invalid or uninterpretable for that purpose. The first question is method validity, not the direction of the effect.
Use language such as ‘the control result must be evaluated against the method’s acceptability criteria before interpreting treatment effects’. Do not invent the exact acceptance threshold for a different sea urchin species or protocol. When a control fails, a report may need to describe the failure and repeat testing under the laboratory’s accredited process. Translation should carry that limitation forward rather than turn an invalid test into a confident environmental conclusion.
Transfer exercise: identify an overnight cohort mix-up
A hatchery starts batch A at 21:15 and batch B at 23:45. A single spreadsheet export lists observations at 00:30 the following morning. At that time batch A has elapsed 3 hours 15 minutes and batch B only 45 minutes. If the export is sorted by clock time but loses its date or batch key, the two microscopy observations can be compared as though their embryos were the same age.
Repair the record by preserving batch ID, insemination timestamp, observation timestamp and elapsed age in one consistent timezone. Batch A’s age is 195 minutes and batch B’s is 45 minutes. If the intended stages depend on elapsed time, their different appearances may be expected. An interpretation that does not reconcile those clocks is not a sound developmental comparison.
Questions readers often ask after the first table
What is the difference between egg recovery and fertilisation rate? Recovery measures the estimated fraction remaining after a handling step; fertilisation rate scores a defined biological marker among examined eggs. Does a high fertilisation score imply high juvenile output? No. Later larval survival, competent development, settlement and nursery survival remain separate stages. Can every donor be evaluated once gametes are pooled? Not without donor-resolved sampling or an appropriate genetic method. Can a fertilisation envelope prove all later development will be normal? No.
Why does dilution notation matter? Because a factor of ten applied as eleven or nine changes the entire estimated starting population. What does n represent? It should be the number of defined independent observations or experimental units, not merely a convenient count of microscope fields. Can a low outcome identify water pollution or poor gamete quality by itself? No. One needs valid controls, design, appropriate assays and supporting evidence before assigning cause.
An audit trail a second reader can actually reconstruct
A completed report should begin with the broodstock and gamete identifiers. Next come collection and processing timestamps, stock volumes, dilution definitions, aliquot volumes and counts, derived total eggs, selection or retention steps, insemination and sampling intervals, fertilisation marker criteria, and later embryo-stage results. Each percentage is followed by the numerator and denominator used, or by an unambiguous reference to those columns. Estimated values should remain distinguished from directly counted observations.
Keep raw source terms beside translations when technical ambiguity remains. Preserve the original unit of measure and original numeric formatting in an audit copy. Verify that every example can be recomputed, that batch boundaries remain intact, and that no later-stage percentage has been relabelled as lifetime survival. The goal is a report that a different aquaculture scientist, laboratory reviewer or translator can read without having to reverse-engineer assumptions from a polished conclusion.
Finally, maintain the limit between documenting established steps and prescribing an operating procedure. A source might describe species-specific sperm dilution, temperature or handling practices under expert supervision. These are contextual research findings, not universally safe recipes. A good educational article teaches a reader how to interpret what was recorded, and where to obtain the actual authorized method for the organism and facility.
How to explain the difference to a non-specialist reader
Imagine a hatchery manager saying, ‘We collected one million eggs, achieved 90% fertilisation, and produced 100,000 juveniles.’ A reader might wrongly conclude that 800,000 fertilised eggs were killed by one identifiable failure. In reality the figures describe several distinct endpoints, usually measured at different times and sometimes using different estimation methods. Between them lie embryonic development, larval rearing, competence, settlement, metamorphosis and nursery survival. Even an exact overall attrition number cannot isolate the stage or cause of losses without intermediate counts. The explanation should therefore identify missing checkpoints and distinguish an arithmetic difference from a biological diagnosis.
A better plain-English rendering is: ‘The initial egg estimate was one million. Under the fertilisation scoring method, approximately nine-tenths of assessed eggs showed the chosen marker. The recorded juvenile output was 100,000. The information provided does not show how much loss occurred at each developmental stage.’ This wording communicates the scale and meaning of the record while avoiding an invented causal story. It remains useful to a non-specialist reader because each stated observation is concrete and the uncertainty is localised.
Why some apparently contradictory studies can both be correct
One paper may report fertilisation percentage at thirty minutes, another normal cleavage at two hours, and a third survival to a swimming larval stage the following day. Even for the same species, those endpoints cannot be directly ranked as if they were three estimates of one common efficiency. The clock origin, water conditions, stock source, number of eggs assessed and morphology used for scoring all matter. A lower later-stage percentage can coexist with a high early fertilisation result without any inconsistency.
When comparing papers, build an evidence table with columns for organism, stage, time since insemination, initial population, scored population, endpoint definition and percentage formula. Identify whether a result is a direct proportion, a conditional survival step, an estimated whole-tank count or a ratio relative to a control. If an abstract suppresses those distinctions, seek the methods section before quoting it. Honest comparisons are often narrower than promotional summaries but more durable and reusable.
Primary sources and nearby learning
NOAA Fisheries: how native sea urchins are spawned and raised in Hawaiʻi
Environment and Climate Change Canada: echinoid fertilisation biological assay