What a postlarval stress-test result actually measures
A shrimp hatchery may supply a document saying that ninety-four percent of postlarvae survived a challenge, while another hatchery reports eighty-eight percent. Those figures can appear comparable at first glance. They are not comparable until we know the animals’ species and developmental stage, the test conditions, exposure duration, number tested, survival definition and whether a non-challenged control was observed at the same time.
A stress test is a controlled observation of response to a specified challenge, not a magical certificate of health. Survival after an acute salinity change tells the reader something about tolerance under that procedure. It does not prove freedom from pathogens, guarantee later pond survival or measure every dimension of larval quality. A useful translation names the challenge and its timing before offering an interpretation of the fraction that lived.
This guide explains how to read published experiments and production quality-control data from first principles. It does not provide an operating protocol for exposing live animals to hazardous chemicals or stressful environments. Test conditions must be approved and conducted by competent personnel with suitable animal welfare, biosecurity and laboratory safeguards. All worked counts in this guide are invented educational examples, not claims about measured hatchery performance.
Translate PL age before comparing any number
PL means postlarva, but the number following it usually identifies days counted from the relevant postlarval transition under the source’s convention. PL2 and PL15 are not interchangeable animals. Developmental changes can affect physiology, feeding, osmoregulation and tolerance to an environmental challenge. If one shipment was tested at PL5 and another at PL12, different survival percentages may reflect age rather than quality.
A 1998 research paper indexed in FAO AGRIS by Samocha and colleagues specifically investigated age-dependent resistance in early Penaeus vannamei postlarvae. That context is crucial. A challenge level that distinguished batches at one age might produce a very different distribution at another. Do not copy an old experimental threshold as a universal contemporary pass mark, particularly for different strains and production systems.
The first entry on a reading worksheet should therefore specify species, PL age, developmental convention and lot history. Also note whether animals were recently transported or fed. It may be tempting to describe a low score as weak larvae, but that conclusion requires distinguishing biological condition from unequal age, handling or exposure.
Survival is a numerator divided by a starting number
The elementary survival calculation is the number that meet the alive criterion at a declared end time divided by the number initially at risk. If 96 of 100 postlarvae are confirmed alive at the endpoint, survival is ninety-six percent. If 48 of 50 are alive, survival is also ninety-six percent, but the sample size and precision differ.
The denominator is not the number recovered alive in the final cup; it is the actual number that entered the test, adjusted only for explicitly justified exclusions. Shrimp that escaped due to a container failure should not be silently counted as biological deaths. Likewise, removing uncertain animals from the denominator without a predefined rule can inflate apparent survival.
Before calculating, reconcile initial count, live count, dead count, missing animals and excluded specimens. A simple ledger might say 100 entered, 93 living, 5 dead and 2 not recovered. Ninety-three percent were observed alive, but the true survival status of two missing animals is uncertain. Recording a ninety-three percent confirmed-alive fraction and the two missing cases is better than inventing their fate.
End time changes the percentage
A result of ninety percent survival at fifteen minutes does not automatically mean ninety percent at one hour. Stress-related mortality can occur after an initial delay or during recovery. Two protocols that use the same environmental challenge but different endpoints measure different outcomes. Always put elapsed time beside the fraction.
Imagine a hypothetical batch of 100 postlarvae. Ninety-eight are alive at fifteen minutes, ninety-six at thirty, ninety-two at sixty and eighty-nine after an additional recovery observation. Those are four valid time-specific observations of the same cohort, not four competing estimates of an unchanging quality trait.
A report that quotes only the best early reading can create a falsely reassuring impression. Equally, quoting only the late endpoint without saying that it includes recovery conditions can mischaracterise the challenge itself. The time course is often more useful than a single percentage because it shows when losses occurred.
Challenge survival and routine survival are different outcomes
Animals may survive normally in the holding tank but die when challenged. The point of a challenge experiment is to expose a difference under a controlled condition, yet that difference must be interpreted within the experiment’s purpose. It cannot directly predict the exact fraction that will survive months later in a commercial pond.
For example, suppose a routine-control group has 99 survivors among 100 animals and a challenge group has 90 among 100. Their observed survival difference is nine percentage points. That comparison describes the groups under the stated conditions. It does not establish nine percent of all later farm deaths were caused by the challenged factor.
A practical translation should separate immediate test tolerance, later nursery performance and harvest survival into distinct metrics. They have different starting populations, durations and causes of loss. A hatchery’s quality assessment can support a decision, but should not be presented as an infallible forecast.
Salinity is the environment, not the outcome
Salinity describes dissolved salt content under a particular measurement convention. A salinity shock describes a change from one exposure condition to another, while salinity tolerance describes an observed biological response under specified circumstances. The words belong to different parts of the experimental sentence and must not be treated as interchangeable.
Suppose animals are held in water described as 28 parts per thousand, then observed in a challenge condition described as 18 parts per thousand. The difference is ten parts per thousand, and the new salinity is eighteen parts per thousand. A report about a ten-part-per-thousand decrease is not the same as a report about exposure to a final salinity of ten parts per thousand.
Because the starting environment matters, give both starting and final values where available. Salinity instruments and reporting units also need attention. The notation ppt is common in older aquaculture papers, while modern measurement descriptions may use practical salinity or other conventions. Do not convert a number merely to make it look familiar without checking the measurement basis.
Rate of change is not the same as endpoint
Two groups can reach the same final salinity using different transition histories. One group might experience a rapid change while another reaches the same endpoint gradually. Even if the final measurement is identical, the exposure trajectories differ and can elicit different physiological responses.
A study of freshwater acclimation indexed in FAO AGRIS by McGraw and Scarpa illustrates why acclimation time, a holding period and ionic composition are meaningful variables. Its findings do not mean a translator should turn an experimental procedure into a universal handling instruction; they show why a number called final salinity is inadequate on its own.
For reading, ask for initial salinity, final salinity, duration of transition, duration held at the endpoint and recovery observation. Also ask whether water differed in its ionic profile. A comparison that preserves these details can help explain seemingly inconsistent outcomes across publications.
Salinity concentration and ionic composition differ
A water sample can have a reported total salinity similar to another while differing in sodium, potassium, calcium, magnesium and chloride composition. Shrimp physiology depends on the ionic environment, not just a total concentration number. Treating salinity alone as a complete description can therefore miss a relevant cause of response.
A hypothetical control and challenge both report a final salt measure of five parts per thousand, but one was made by diluting seawater and the other from a different mineral mixture. The equal headline number does not guarantee equal ionic conditions. If survival differs, it would be premature to attribute the whole effect to the salinity level.
Where available, translate ionic measurements with species, concentration units and laboratory methods intact. If detailed chemistry is missing, state that limitation. For decisions about animal health or water chemistry, consult qualified specialists rather than using a single salinity reading as a comprehensive diagnostic.
Temperature challenges require exact timing
Temperature influences metabolism and animal response, and a thermal challenge is defined by both temperature history and observation schedule. The final temperature alone does not reveal the duration of exposure or how rapidly the water changed. Nor does one study’s experimental temperature define a universally safe operating threshold for all postlarvae.
Consider two published comparisons in which one group was observed for fifteen minutes at a test temperature and another for two hours. Even if both report a percentage alive, their observation windows are not equivalent. A lower percentage in the longer experiment may partly reflect additional time at risk rather than intrinsically poorer animals.
Temperature should be documented as a measured value with appropriate precision, alongside ambient conditions and equipment calibration where known. Avoid converting a controlled laboratory challenge into advice to reproduce a harmful exposure. The purpose here is accurate reading of the evidence and its limitations.
Chemical challenges and professional safeguards
Some historical studies have used chemicals such as formalin to investigate larval stress tolerance. A published concentration-response curve can be interpreted mathematically without reenacting the exposure. Formaldehyde-containing solutions present chemical and animal-welfare hazards, and any controlled laboratory use must be confined to qualified operators under applicable requirements.
A translation should identify whether the source measured chemical concentration in the test solution, exposure duration, mortality endpoint and recovery observation. It should never remove the experimental context and issue an unsupported directive such as expose every shipment to this dose. A historical LC50 describes a study’s estimated response under a declared design, not a recommended dose.
Where a reader’s real goal is to evaluate hatchery quality, compare established, validated and welfare-conscious screening methods with veterinary guidance. A single stressful test is not a substitute for pathogen screening, husbandry records, microscopic assessment or appropriate longer-term performance data.
Controls help locate ordinary background loss
A challenge survival value is more interpretable when compared with an appropriate control from the same source lot and observation period. Controls help reveal mortality or handling effects that occur even without the target challenge. They also help identify whether an experimental setup itself caused difficulty for animals.
Suppose a control starts with eighty animals and ends with seventy-eight alive, or 97.5 percent, while a challenged group also starts with eighty and ends with sixty-eight alive, or eighty-five percent. The observed absolute difference is 12.5 percentage points. That is a comparison, not an assertion that a specific biological mechanism caused every additional death.
If the control itself has unexpectedly poor survival, a reader should be cautious about assigning meaning to the treatment group. Record control conditions, random assignment and whether the same counting criteria were used. A missing control limits causal interpretation even where the challenge result has a precise-looking number.
Percentage points and relative percentages
A survival difference can be described in percentage points or as a relative percentage, but the two must not be confused. If group A survives at ninety percent and group B at eighty percent, the absolute difference is ten percentage points. Relative to B, A’s observed survival proportion is 12.5 percent higher, calculated as ten divided by eighty.
The reverse comparison also has a different relative base. B’s survival is about 11.1 percent lower relative to A, calculated as ten divided by ninety. None of these is simply a ten percent improvement in every sense. When translating results, retain the reference group and distinguish changes in proportions from changes in counts.
For practical hatchery discussions, percentages points are usually clearer when comparing two survival percentages. A ten-point difference among small samples also carries uncertainty. The visual size of a percentage must not obscure whether twenty or two thousand animals were observed.
Sample size determines granularity
When twenty postlarvae are tested, one death changes the observed survival rate by five percentage points. When 200 are tested, one death changes it by half a point. The same displayed ninety-five percent may therefore arise from nineteen of twenty animals or 190 of 200, with different statistical precision.
A result of one hundred percent in a small sample does not prove the underlying risk of mortality is zero. It means no deaths were observed among those tested under the reported conditions. Likewise, a single death in ten animals creates an apparent ten-percent mortality estimate that is much less precise than a similar rate observed across many independently sampled animals.
Keep numerator and denominator attached to every percentage. In dashboards, display 19/20 beside 95%, not just a green or red symbol. That small formatting decision makes uncertainty visible without requiring a reader to understand formal statistical theory.
Tank replication versus counting many individuals
Postlarvae within the same test vessel share water chemistry, temperature, oxygen, handling and other conditions. Their fates may be correlated. Counting hundreds of animals in one vessel does not necessarily provide the same experimental replication as several independently assigned vessels under each treatment.
Suppose one control tank contains 200 larvae and one challenge tank contains 200. A spreadsheet could claim 400 observations, but the treatment itself was applied once in each tank. It is impossible to separate a treatment effect fully from an accidental difference between those two tanks without additional replication.
For a careful scientific translation, identify the experimental unit and replicate structure. Report both animal counts and independent vessel counts. The difference is important if a paper quotes a p-value or confidence interval that depends on assumptions about independence.
Define alive, dead and moribund
Survival requires operational criteria. An animal that moves when gently observed, an animal that is immobile but later recovers, and an animal confirmed dead are not necessarily the same categories. If studies use different criteria, their percentages may disagree for a measurement reason rather than a biological difference.
Suppose a hypothetical endpoint includes eighty active postlarvae, eight weakly responsive animals and twelve confirmed dead out of 100. A strict active-only proportion is eighty percent, whereas an alive-including-weak fraction could be eighty-eight percent, if the classification protocol supports it. Neither should be labelled simply survival without revealing the rule.
Observation methods, magnification, water movement and time after disturbance can affect categorisation. The correct approach is to report the source’s criteria and provide categories where possible, not adjust classifications after seeing which result is more favourable.
Disappearance and handling loss are not necessarily deaths
Some animals can be missing at the endpoint because of equipment, transfer, adhesion or counting errors. A survival calculation that simply subtracts the missing from the initial count may overstate mortality. Conversely, deleting the missing animals from the denominator may inflate the fraction alive.
If 100 postlarvae enter a test, 88 are confirmed alive, eight confirmed dead and four are unaccounted for, observed confirmed-alive proportion is eighty-eight percent. Among the 96 classified animals, the conditional live fraction is about 91.7 percent, but that figure excludes the four unknown outcomes and cannot replace the original cohort result.
Show both the accounting gap and any conditional percentage rather than forcing all animals into a binary category. Reconcile transfers, containers and sampling equipment before drawing a biological inference. Missingness can itself be an important quality-control warning.
LC50 means a modelled median lethal condition
LC50 is the concentration at which an experiment or fitted concentration-response model estimates fifty percent mortality over a specified duration and under stated conditions. In salinity research, authors may describe a median lethal salinity or a difference from starting salinity. These are not necessarily expressed or interpreted in the same direction as a toxic chemical concentration.
The meaning of LC50 depends on the endpoint and experimental design. A two-hour LC50 cannot be compared directly with a twenty-four-hour LC50 as if time were irrelevant. An LC50 estimated from a fitted curve is not necessarily an exact concentration directly observed to kill exactly half the animals in one vessel.
When translating, preserve the units, exposure duration, species and PL stage, plus confidence intervals if given. Do not call LC50 a safe dose, a no-effect level or a regulatory limit. It is a summary of an experimental mortality relationship and should be interpreted within that relationship.
Survival and mortality are complements only with complete classification
If every animal can be classified as either alive or dead at the same endpoint, survival plus mortality equals one hundred percent. That elementary identity is a useful audit check. However, missing animals, exclusions, recovery states and differently timed observations can break the simple two-category accounting.
A report stating ninety-two percent survival and ten percent mortality for the same defined population should trigger a denominator check. The values may come from different populations or rounding conventions, but they cannot both be exact complements on a complete single cohort. Do not automatically alter one number until the underlying records are reviewed.
If survival is counted at sixty minutes and mortality after twenty-four-hour recovery, the two values describe different periods. Put the endpoint beside each value, and explicitly identify any individuals whose status is uncertain.
A full hypothetical challenge ledger
Imagine that four replicate vessels each begin with fifty PL12 whiteleg shrimp. At a declared observation endpoint, the live counts are forty-seven, forty-six, forty-eight and forty-five. Total starting animals equal 200, live animals 186 and confirmed dead animals fourteen, assuming no missing classifications.
Pooled observed survival is 186 divided by 200, or ninety-three percent. Vessel rates are ninety-four, ninety-two, ninety-six and ninety percent. The arithmetic mean of those four equal-sized vessel rates is also ninety-three percent. That agreement occurs because all vessels started with equal denominators; it should not be assumed for unequal starting numbers.
The vessel-to-vessel variation tells us something that a pooled percentage alone omits. If every vessel had a different survival outcome, the results need a replicate-level analysis and a record of environmental conditions. This example only demonstrates arithmetic; it does not assess whether ninety-three percent is good for a particular test.
Unequal vessel sizes and pooled percentages
In a fictional experiment, one vessel starts with twenty animals and ends with eighteen alive, while another starts with eighty and ends with sixty-four. The vessel survival rates are ninety and eighty percent. Their arithmetic mean is eighty-five percent, but pooled survival is eighty-two percent because eighty-two of the 100 total animals survived.
The first summary gives each vessel equal weight; the second gives each animal equal weight. Which one to report depends on the research question and analysis plan. If vessels were independently assigned, the replicate-level pattern matters, and a purely pooled calculation can mask variability among experimental units.
A translator should give the denominators in prose and avoid calling eighty-five percent or eighty-two percent the unique correct overall result without context. The arithmetic does not supply the scientific reason for preferring one summary.
An apparent perfect result
Suppose all twenty tested postlarvae survive a sixty-minute challenge. The observed survival is twenty out of twenty, or one hundred percent. That is a true description of the observed sample, but it does not establish that every future animal of that strain will survive the same challenge.
If a second study tests 400 animals and finds 392 survivors, its observed survival is ninety-eight percent. The small study’s one hundred percent may sound superior, yet the evidence is far too limited to rank the two production systems confidently without matched conditions and uncertainty analysis.
Use language such as no mortality observed among twenty tested animals rather than risk-free or guaranteed. Reporting the count immediately guards against a misleading reading of a perfect percentage.
Why the same batch can receive different scores
Animals within a hatchery batch may vary in size, developmental stage, feeding status, condition and recent handling. Subsamples drawn from different positions may not be identical. Test vessels can differ slightly in water quality, temperature or counting quality despite efforts to standardise them.
If one sample yields ninety-four percent and a repeat yields eighty-eight percent, first compare sample sizes, classification rules, age, exposure and procedures. The difference may reflect real heterogeneity, sampling variation or procedural changes. A single pair of percentages cannot determine the cause.
The most useful report records selection method, source vessel, test date, replicate IDs and subsequent observations. A discrepancy becomes a scientific question rather than a reason to choose whichever result supports a desired purchase decision.
Recovery observations matter
Immediate survival under a challenge and survival after a recovery interval are related but distinct. Some postlarvae that appear stressed may recover; others may initially appear normal but decline later. A study that measures both should name and preserve both endpoint conditions.
Suppose 95 of 100 are alive at the end of the exposure, and 91 are alive after recovery. Exposure-end survival is ninety-five percent. If the same initial cohort is followed without new animals, overall survival through recovery is ninety-one percent. Conditional recovery survival among the 95 alive at exposure end is approximately 95.8 percent, provided the records support that sequence.
The two denominators are not substitutes. The conditional figure answers what fraction of exposure survivors persisted through recovery, while the overall figure answers what fraction of the original cohort remained alive. A precise translation makes that difference visible.
Why predictive correlations are not guarantees
Some published studies report that challenge survival was correlated with later larval or postlarval survival under their experimental conditions. A correlation can provide useful evidence of association, but it does not prove that a stress-test score predicts future survival accurately for every animal, region, strain or farm.
A reported correlation coefficient, such as 0.42 in an observational relationship, is not a statement that forty-two percent of postlarvae will survive. Nor is it a success rate. The coefficient quantifies the strength and direction of a relationship in the analysed data under assumptions about the variables and method.
Where a correlation is used for screening, consider validation on comparable animals, calibration, uncertainty, false positives and false negatives in practical decisions. Do not turn statistical association into deterministic advice about stocking success.
Keep diagnostic PCR separate
Stress-test survival describes an observed response to a challenge. PCR describes detection of specified nucleic-acid targets under validated assay conditions. A shrimp can survive a stress test while carrying an infection, and a molecular-positive shrimp may appear outwardly normal at the time of sampling. One result does not cancel the other.
A hatchery quality assessment may combine multiple independent lines of evidence, including developmental observations, health screening, husbandry history and transport performance. A single survival percentage should not be translated into a pathogen-free certificate. Similarly, a PCR result needs its own sample type, assay interpretation and laboratory quality controls.
The vocabulary matters: tolerance, survival, infection status and later performance are four different ideas. When these labels are kept distinct, a producer can make better decisions without overstating what any one test establishes.
Beware a pass threshold without validation
A supplier may call a batch acceptable if observed survival exceeds a chosen percentage. The threshold may be part of a contractual specification, local procedure or research design; it is not automatically a universal biological law. Ask who selected it, for which age and species, using which validated test and for which decision.
An eighty-percent pass threshold applied to a small sample can classify batches differently after one animal’s outcome changes. For fifty animals, forty survivors pass and thirty-nine fail if the rule is at least eighty percent. That one-animal boundary may be operationally required, but the underlying biological difference is not magically sharp.
A transparent report identifies the chosen rule as a rule, not a measured truth about quality. It also names what happens when sample counts are incomplete, tests are invalid or uncertainty straddles the decision boundary.
Ethical and welfare considerations in interpretation
An educational guide can explain experimental stress-response statistics without recommending that animals be subjected to dangerous exposure. Researchers and hatchery professionals have obligations to minimise unnecessary suffering, apply suitable oversight and use scientifically justified methods. Chemical challenges deserve particular caution because they may introduce hazards to workers and animals.
A poor score can lead to the humane improvement of conditions, better sampling or veterinary consultation rather than simply more severe testing. A strong quality system seeks valid information with the least harm needed to answer its question. Report whether relevant ethical and procedural standards were documented when evaluating a published study.
The absence of an ethical statement in an older paper should not be filled in by assumption. Distinguish historical description from current recommendation and direct operational decisions to qualified teams following applicable rules.
How to investigate conflicting test reports
Start with the biological identity and stage. Then compare starting conditions, final exposure, change rate, hold duration, sampling method, live/dead definitions, number of independent replicate vessels and control performance. This checklist often reveals why two figures that look contradictory are not measurements of the same question.
If those details match, examine raw counts and uncertainty. An apparent five-point difference might be a small-sample fluctuation. If repeated tests with independent replication still diverge, consider lot heterogeneity, storage or transport history and assay implementation. Do not assign a mechanism without supporting measurements.
A good scientific translation includes a sentence explaining what remains unresolved. The goal is not to erase uncertainty; it is to give readers a reliable map of what each result can and cannot say.
Worked example: a control and a challenge
A fictional study uses five independent control vessels and five challenge vessels, with forty animals in each. Across controls, 196 of 200 are confirmed alive. Across challenged vessels, 174 of 200 are alive. Calculate the observed survival and absolute difference, then identify what additional information is needed for a formal treatment comparison.
Control survival is ninety-eight percent and challenge survival eighty-seven percent. The difference is eleven percentage points. Observed deaths total four and twenty-six respectively if no animals are missing. The counts make a descriptive comparison possible, but the reader also needs survival for each replicate, assignment scheme, exposure confirmation and a suitable statistical analysis.
An interpretation that says the challenge caused exactly eleven percent additional deaths across all future batches would exceed the evidence. The demonstration shows a sample difference under controlled conditions, not a universal causal magnitude.
Worked example: survivors measured twice
One hundred and twenty animals start a hypothetical challenge. At the end of exposure 114 remain alive, and after recovery 108 remain alive. Calculate survival at the two endpoints and conditional survival during recovery. Show the denominators in a sentence that could be used in a scientific summary.
Exposure-end survival is 114 divided by 120, or ninety-five percent. Survival through recovery is 108 divided by 120, or ninety percent. Conditional recovery survival is 108 divided by 114, or approximately 94.7 percent. These figures describe the same cohort with different observation windows and risk sets.
A suitable summary is: of 120 postlarvae exposed, 114 were alive at the immediate endpoint and 108 remained alive after recovery; six of the 114 immediate survivors were no longer alive at the later observation. The distinction prevents a reader from confusing conditional recovery with survival of the entire starting cohort.
Worked example: age-confounded comparison
A fictional supplier provides survival results of ninety-six percent for PL15 animals and eighty-eight percent for PL5 animals. Does that establish that the first source’s animals are higher quality? No. Age differs substantially, so a valid quality comparison would require age-matched testing or a justified model accounting for age.
The observed eight-percentage-point difference can be reported, but its interpretation must say the conditions were not matched. Research into postlarval tolerance shows that age-related physiological changes can matter. A higher score could reflect maturity as well as other conditions.
A stronger design would sample comparable developmental stages with independent replication, defined exposure and controls. The purpose of the example is not to recommend any challenge procedure, but to teach the habit of checking whether percentages refer to biologically comparable animals.
Worked example: saline endpoint versus change
A hypothetical paper states that a test began at 30 ppt and ended at 12 ppt. Another paper reports a decrease of 12 ppt from 30 ppt. Calculate both final conditions. The first is 12 ppt. The second is 30 minus 12, or 18 ppt. The figures describe different exposures despite sharing the number twelve.
If a translation drops from or to, the experiment changes meaning. This error is especially serious in reports that rely on a critical environmental transition. Preserve original units, starting values and prepositions rather than smoothing the text into vague wording.
As a checking technique, rewrite the statement as starting value, operation and ending value before interpreting any survival difference. If the numbers cannot be reconciled, consult the original methods and leave the ambiguity visible.
Worked example: missing animals
A hypothetical vessel begins with sixty postlarvae; fifty-two are observed alive at the endpoint, five dead and three missing. The confirmed-alive proportion of the initial group is 52 divided by 60, or approximately 86.7 percent. Among the 57 classified animals, 52 divided by 57 is approximately 91.2 percent.
The second percentage is not equivalent to complete-cohort survival. It describes the classified subset and could overestimate the true fraction alive if missing animals died. Counting all three missing animals as dead without evidence could instead overstate mortality.
A faithful summary reports fifty-two live, five dead, three unaccounted for, and the resulting qualification. Repair the counting and handling record before relying on the number for a pass-fail decision.
Worked example: percentage-point threshold
A policy says a fictional batch passes when at least ninety percent of fifty tested animals are alive. Determine the minimum number of survivors and examine how one outcome affects the decision. Forty-five survivors yield ninety percent and pass; forty-four yield eighty-eight percent and fail.
The rule is arithmetically straightforward. But it is an administrative classification built on a particular sample size and protocol. It does not prove a fundamental discontinuity in biological quality between two otherwise similar lots. A system should explain how invalid tests, missing observations and replicate disagreements are handled.
An accurate translation may state meets the stated acceptance criterion rather than guaranteed suitable for any pond. The narrower wording is the one supported by evidence.
Independent problem: compare two experiments
Experiment A reports 144 survivors among 150 PL10 animals after a defined observation. Experiment B reports 91 survivors among 100 PL10 animals, but its exposure lasted twice as long. Calculate the proportions and explain whether the higher percentage can fairly be ranked as a superior quality result.
A gives ninety-six percent and B gives ninety-one percent. The difference is five percentage points, but the exposure duration differs. Without a matched protocol, the numerical ranking cannot isolate quality. A sensible answer would request the time-specific survival curves, replicate information and controls.
The lesson is that arithmetic is necessary but not sufficient for biological interpretation. Two correct percentages may answer two different experimental questions.
Independent problem: weighted versus unweighted
Three hypothetical replicate vessels have starting counts of twenty, thirty and fifty, with eighteen, twenty-four and forty-five survivors respectively. Find the individual rates, their simple average and pooled survival. Identify why the two combined values need not match.
The individual rates are ninety, eighty and ninety percent. The unweighted mean is approximately 86.7 percent. Pooled survival is 87 survivors divided by 100 animals, or eighty-seven percent. The near agreement is coincidental; the summaries use different weights.
If the experiment regards each independently assigned vessel as a unit, replicate-based analysis should not be replaced uncritically by a pooled percentage. Record the experimental question before choosing an aggregate.
Independent problem: conditional recovery
A cohort of eighty animals has seventy-six survivors at exposure end and seventy-two after recovery. Calculate survival through exposure, overall survival through recovery and conditional recovery among immediate survivors. Then write the three rates as complete sentences.
They are 76/80 = ninety-five percent, 72/80 = ninety percent, and 72/76 = approximately 94.7 percent. The first two use the original cohort denominator, while the third uses only the animals alive at the end of exposure.
A reader who describes the last value simply as final survival would alter the study’s meaning. A reliable translation states when and among whom survival was measured.
Independent problem: classify the statement
A hypothetical procurement document says ninety-five percent survived, so this stock is disease free and will achieve ninety-five percent harvest survival. Identify two unsupported leaps. First, a challenge survival result is not a molecular or clinical declaration of pathogen status. Second, a short observation does not determine months-long harvest survival across variable pond environments.
A repaired statement would say that ninety-five percent of the sampled postlarvae met the survival criterion under the stated test conditions. If the document also contains separate valid pathogen surveillance and production data, those should be reported independently with their own methods.
This exercise is about translating claims without smuggling additional conclusions into a numerical sentence.
FAQ: What does a high stress-test score establish?
It establishes that a high proportion of the tested animals met the specified survival criterion under a particular challenge and observation window. Depending on design, that may support an assessment of tolerance or batch condition. It does not by itself establish freedom from pathogens, a universal quality grade or a guaranteed future harvest outcome.
Ask for the actual numbers, species, developmental stage, testing method and controls before relying on the result. A percentage without context is incomplete evidence.
FAQ: Can two tests with the same salinity be compared?
Not automatically. They may differ in starting salinity, rate of change, holding time, water ionic composition, temperature, PL stage and observation endpoint. Equal final salinity does not establish equal treatment histories.
If those factors are not reported, the safest explanation is that the results are not directly comparable on the information available. Avoid filling the gap with an invented equivalence.
FAQ: Does one hundred percent survival mean no risk?
No. It means no deaths were observed among the tested animals at the stated endpoint. The result may arise from a small sample, an insufficiently discriminating challenge, strong animals or other features of the test. It is not proof of zero underlying mortality probability.
Write the numerator and denominator together. Twenty of twenty is a different evidentiary basis from two thousand of two thousand, even though both display one hundred percent.
FAQ: Is LC50 an acceptable operational dose?
No. LC50 is an experimental median lethal condition with duration and design attached; it should never be translated as safe, recommended or no-effect exposure. With chemical stressors, laboratory handling and animal welfare require specialist oversight.
The educational task is to interpret the relationship and distinguish LC50 from threshold claims, not to prescribe a procedure.
FAQ: Why can sample sizes change the decision?
A single animal represents a different fraction of a sample depending on its size. One death in twenty corresponds to five percentage points, while one in 200 corresponds to half a point. A small sample’s apparent quality ranking can therefore be unstable.
Inspect the counts and the facility’s predefined decision rule. A published percentage that omits its denominator should be treated as incomplete.
FAQ: What if a control group also suffers loss?
Then the challenge result is harder to attribute solely to the challenge. Background mortality, handling or vessel conditions may contribute. The appropriate interpretation compares matched control and challenged groups and investigates whether the controls met the study’s validity criteria.
Do not automatically subtract a control percentage from a test percentage without understanding the relevant design and statistical method. Show the observed rates first.
FAQ: What should an honest one-sentence summary say?
A suitable fictional summary is: in a defined sixty-minute test of fifty PL12 whiteleg shrimp, forty-six met the live criterion at the endpoint, corresponding to ninety-two percent observed survival; the result describes this test rather than future pond performance.
This sentence retains the biological group, duration, count and limitation. Its specificity makes it more useful than a vague statement that the larvae passed every quality check.
Primary research and related reading
FAO AGRIS: Samocha et al. 1998, simple stress test for P. vannamei postlarvae
FAO AGRIS: criteria for assessing larval and postlarval quality
FAO AGRIS: McGraw and Scarpa, acclimation rate and ion composition
FAO AGRIS: Sumarwan et al. 2023, comparisons of stress responses
eduKateSG: shrimp postlarvae salinity acclimation and stocking recovery
eduKateSG: aquatic-animal PCR sensitivity and surveillance