Mud crab hatchery reports become confusing when a count appears to improve while the animals are actually disappearing. A tank can show fewer zoeae today because larvae died, because some have metamorphosed into megalopae, because water was exchanged, or because a sample represented the tank poorly. To translate the report properly, first find the stage, the denominator and the time of measurement. Those three details determine what the number means.
This guide follows mud crabs of the genus Scylla from zoea through megalopa to the first juvenile crab instar. It is for readers who must explain hatchery records in plain English without turning a developmental transition into a survival claim. The examples are invented teaching calculations, not observations about a commercial hatchery, and they deliberately keep biological stage recognition separate from arithmetic.
The most useful question is not simply how many larvae are present. It is how many animals entered a defined stage, how many remained or advanced by a defined endpoint, which departures were recorded, and whether the two counts use comparable sampling methods. Only then can a percentage be assigned a reliable name.
Start with a stage map, not a single survival percentage
Mud crabs undergo a series of zoeal stages, often marked Z1 to Z5, before the megalopa stage and the first juvenile crab stage, commonly C1. The transitions are biological events, not just calendar labels. Larvae are planktonic early in the cycle, whereas megalopae begin taking on a more benthic existence. The shift changes feeding, habitat use, vulnerability to cannibalism and how a sample is collected.
A stage sheet might say “Z5 dominant,” “megalopa first seen,” or “C1 recovered.” None of these statements is synonymous with “all animals reached that stage.” Dominant could mean most of a microscope subsample, not most of a tank. First seen identifies a detection time, not a complete cohort transition. Recovered can mean collected in the harvest operation, not every animal present. Translation should preserve those narrower claims.
The literature is also not a universal timetable. SEAFDEC hatchery work on Scylla serrata has documented zoea-to-megalopa and megalopa-to-crab development over specified culture conditions, but a reader should not transfer those days mechanically to a different species, origin, temperature, salinity or protocol. A count sheet records the actual cohort; a published schedule provides background expectations and possible reasons to inspect unusual timing.
Read the unit attached to a count
“Thirty larvae” could mean thirty organisms in a one-millilitre aliquot, thirty in a ten-millilitre cup, thirty per litre after a conversion, or thirty retained in a net. These are four very different statements. If a laboratory writes 30 individuals/mL, multiplying by 1,000 gives 30,000 individuals/L, assuming the units really are millilitres of the water sample and the sample is representative.
When the tank has a known effective volume, an estimated population follows by multiplying concentration by volume in the same units. For example, 20 larvae/L in 2,000 L yields an estimate of 40,000 larvae. A calculator can obtain this product instantly, but it cannot decide whether the water was mixed, whether the larvae were concentrated near light or shelter, or whether tank volume already excludes a recent drawdown.
Avoid translating “per sample” as “per litre” without a declared sample volume. Similarly, “counted 60” means an observed count, whereas “estimated 60,000” is an extrapolation with uncertainty. The report should label those as distinct quantities. A cautious caption might read: “Mean observed concentration in three mixed subsamples, extrapolated to current working tank volume.” That sentence says more than a bare large number.
Reconstruct the exact sampling operation
Sampling usually has at least five steps: establish mixing conditions, choose a location and depth, collect a known water volume, remove or concentrate a defined part of it, and count under a microscope or with another validated procedure. Translation should keep track of each dilution or concentration factor. A small overlooked factor can be more consequential than the last digit displayed on the result.
Suppose an operator filters 10 L of tank water into 100 mL concentrate, mixes the concentrate, and counts 2 mL containing 160 larvae. The aliquot represents 2/100 of the concentrate, so estimated larvae in the full concentrate equal 160 times 50, or 8,000. Dividing by the original 10 L means 800 larvae/L. Calling the observed microscope count “800 larvae/L” without showing the intermediate factor conceals what happened.
The estimate assumes all larvae from the ten-litre original sample were retained through concentration and were evenly distributed in the final 100 mL. If sieve retention is incomplete, or if animals settle before pipetting, the estimate can be biased. A recovery experiment or replicate aliquots can reveal some of that problem. A translation should therefore say “estimated concentration from concentrated sample,” not “exact tank abundance.”
Worked example: calculate hatch concentration with units
Imagine a hatch tank of 1,500 L. The crew collects three one-litre samples after standard mixing and counts 420, 460 and 440 Z1 larvae per litre. The arithmetic mean is (420 + 460 + 440)/3 = 440 larvae/L. The simple estimated inventory is 440 larvae/L times 1,500 L, or 660,000 Z1 larvae. This is an estimate of animals present at the moment of sampling, not necessarily the number of eggs initially attached to the female.
The sample standard deviation of the three concentrations is 20 larvae/L. This illustrates variation among the three samples but does not magically capture every source of systematic error. If the tank has a poorly mixed corner, the three samples may agree with one another and still miss it. The original report should say whether sampling locations and volumes were predefined and whether the entire tank was mixed in a comparable way.
Suppose the water level then falls and the next working volume is 1,200 L. A concentration of 500 larvae/L at the second time represents approximately 600,000 larvae, not 750,000. The concentration has risen while the inventory has fallen. Without the volume adjustment, someone might claim population growth even though zoeal larvae did not reproduce inside the tank. This is a basic mass-balance check that any translated comparison should pass.
Distinguish hatch rate from larval recovery
A hatch percentage needs a denominator tied to eggs or estimated potential hatch, while a recovery percentage may use estimated larvae present at an earlier time. If a team can estimate viable eggs and newly hatched Z1 larvae, it can define a hatch-based ratio. If only post-hatch larval counts exist, a claim about hatch success cannot be reconstructed from those records alone.
For example, if 800,000 viable eggs were estimated by a stated method and 600,000 Z1 larvae were independently estimated after hatching, the illustrative hatch-to-Z1 recovery fraction is 600,000/800,000 = 75%. Yet that fraction combines biological hatching with losses and sampling between the two measurements. Calling it a true hatching rate would require a method that separates those events. The denominator can be numerically correct but conceptually wrong.
In another report, “600,000 seed produced” may refer to early zoeae, harvestable megalopae, or juveniles depending on the facility’s shorthand. Translate the label “seed” with a parenthetical stage, sample date and method. The reader must know whether this is an initial input, an intermediate stage or a transferable output before drawing conclusions about productivity.
What an asynchronous cohort looks like
Individual larvae do not moult on the same minute. One tank can contain Z4, Z5 and megalopae on the same day. A stage distribution is therefore often more informative than a single stage heading. A sample showing 15% Z4, 55% Z5 and 30% megalopa says the cohort is mixed at that instant; it does not say 30% of the original cohort survived to megalopa unless the numerator and original denominator were counted appropriately.
Suppose 200 organisms are classified in a representative microscope sample: 30 Z4, 110 Z5 and 60 megalopae. Their sample proportions are 15%, 55% and 30%. The percentages sum to 100% because each observed organism was assigned to one mutually exclusive class. If damaged or unclassifiable organisms were excluded, the denominator must be reported: “60/200 classified live larvae,” not “60/215 observed objects.”
Mixed stages matter operationally as well as statistically. The early metamorphosing megalopae may interact differently with larvae that remain Z5. A stage-specific response might be to separate groups, adjust refuge or assess feed. Those are management decisions requiring a validated site protocol, not commands that follow from one sample proportion. A translation should distinguish an observed stage mix from a justified intervention.
From Z5 to megalopa: identify the correct transition
A frequent source of confusion is the “megalopa yield” calculation. Does the numerator represent all megalopae observed over several days, the megalopae currently in a tank, or the cumulative unique megalopae harvested and transferred? If one animal is counted in the tank on Monday and again on Tuesday, adding the daily estimates double-counts that animal. A cumulative harvest ledger should add unique removals, not repeated snapshots.
Consider a tank with an initial estimated 100,000 Z1 larvae. During metamorphosis, the crew harvests and transfers 9,000 megalopae on day A and 5,000 on day B, with records confirming separate removal batches. An illustrative cumulative megalopa recovery is (9,000 + 5,000)/100,000 = 14%. If another 2,000 are estimated still in the original tank at close of the defined window, the inventory-accounting figure could be 16%, but only if transfer losses and later double-counting are ruled out.
Those are not the same as “percentage metamorphosed” because some animals may have become megalopae and died before a count, escaped the sampling frame, or remained undetected. Prefer the precise wording “observed and recovered as megalopae” when that is what the data show. A biological claim should not exceed the reach of the counting system.
From megalopa to C1: do not reset the denominator silently
Once megalopae are transferred into a nursery, the next ratio often uses transferred megalopae as its denominator. If 14,000 megalopae are stocked and 7,000 C1 juveniles are recovered, nursery conversion or C1 recovery is 50% under that specified endpoint. The overall Z1-to-C1 recovery from an original cohort of 100,000 would be 7%, assuming cohorts have not been mixed and no C1 animals entered from another source.
A report may show “50% survival” beside “7% overall yield.” Neither value contradicts the other if the stages and denominators are stated. The first measures performance of a narrower interval; the second connects the final output to the original input. Compare two hatcheries fairly only if they use the same segment and endpoint or provide enough detail to reconstruct both.
If megalopae were transferred to several nursery units, use a cohort ledger. For example, tank N1 receives 8,000 and returns 4,800 C1; N2 receives 6,000 and returns 2,200 C1. The pooled recovery is (4,800 + 2,200)/(8,000 + 6,000) = 50%, not the unweighted average of the separate percentages. Here N1 is 60% and N2 is about 36.7%; their simple mean is about 48.3%, which gives equal weight to unequal populations.
Build a cohort ledger that can be audited
Give each spawn or hatch a stable identifier before combining data across tanks. A minimum ledger records broodstock or spawn reference, hatch date, species identification, starting Z1 estimate and method, every transfer with source and destination, mortality removal counts where observable, water exchanges affecting inventory, stage assessments and final recovery by unit. It should also explain whether counts are direct or estimated.
Suppose cohort M24 starts as 120,000 estimated Z1 larvae, transfers 18,000 unique megalopae into two nurseries, and yields 9,000 C1 juveniles. Label the first stage recovery 15%, nursery recovery 50% and end-to-end recovery 7.5%. In this simple chain, 0.15 times 0.50 = 0.075, which independently checks the overall ratio. The multiplication works only because the numerators and denominators describe the same animals and transitions.
The ledger must not treat every disappearance as a recorded death. Crabs can hide, cling to structures or be omitted by a counting procedure; a discrepancy is an unaccounted difference until evidence establishes its cause. Use “inventory difference,” “unobserved loss” or “unrecovered individuals” when appropriate. A translated report that turns a missing animal into a diagnosed disease death is adding an observation nobody made.
Decide what counts as a verified stage
Stage labels should be assigned using visible anatomical and behavioural criteria from a competent identification guide, not merely the day number on the wall calendar. If the staff use a shorthand code, provide its local definition. A postlarval stage recorded as “M2” might mean a two-day-old megalopa rather than a second distinct megalopal instar; these coding systems must not be assumed equivalent.
A first crab may be recorded as C1, crab instar one, or first juvenile. These terms often aim at the same transition, but a specific study may have a different count or sampling endpoint. Translate with the local protocol and a short explanation: “C1, the first crab-shaped instar after megalopa.” If a classifier cannot distinguish a stage reliably, report “uncertain stage” rather than fabricating precision.
Photographs with scale bars, preserved reference specimens and a written classification key can improve consistency across technicians. A stage record should also say whether recently moulted, injured or dead organisms were included in the counted set. Counts of alive larvae and all larval bodies answer different questions. A percentage cannot be interpreted when the inclusion rule changes between days.
Why feed and water records belong next to counts
Larval survival is not explained by counts alone. The FAO mud crab aquaculture manual describes the importance of monitoring larval stage, appearance, density, feeding and water conditions. These observations provide hypotheses about performance, but do not establish a cause merely through temporal association. If the cohort declines after a temperature shift, investigate possible multiple explanations before blaming the temperature entry.
A useful combined table records sample time, stage mix, live-feed density, additions of feed, salinity, temperature, dissolved oxygen when measured, water exchange, visible behaviour and count method. Convert all volumes to stated units and preserve the original instrument precision. Two salinity observations at different times do not measure the same exposure as a continuous record, and a missing observation must be kept missing.
The purpose of this context is to make numbers interpretable and prompt responsible follow-up. For example, rising megalopa proportion alongside a falling total inventory may reflect healthy metamorphosis accompanied by losses. The stage distribution alone cannot distinguish cannibalism from sampling bias. Do not turn descriptive co-occurrence into an experimentally proven mechanism.
A practical reading order for a crowded hatchery sheet
Begin with the date, local time and cohort identifier. Next locate the tank’s effective water volume at the same time as the concentration sample. Read the precise unit, sample and concentration steps, and whether the numerator comprises live organisms. Then decode the biological stages and identify transfers or harvests between observations. Only after reconstructing those parts should you calculate changes.
Mark every calculated quantity explicitly. A technician may write “Z1: 400 per litre” and “population: 800,000”; the second may be a correct expansion from a 2,000-litre tank, but it is not a separate independent observation. If someone later averages the raw concentration and estimated population as if both were replicate measurements, the result becomes meaningless. Translation can protect readers by labelling derived values.
End with a claim no stronger than the data support: “The reported Z1 concentration was 400/L; at the recorded tank volume the estimated inventory was 800,000, subject to mixing and sampling uncertainty.” This is both clearer and more defensible than “800,000 healthy larvae were produced.” It preserves a boundary between a numerical estimate and a biological or commercial conclusion.
Why one-day survival is not necessarily a hatchery survival rate
A daily percentage is meaningful only after specifying the population entering that day. If there were 50,000 animals on Monday and 45,000 on Tuesday, a simple apparent retention ratio is 90%. But if 6,000 animals were intentionally transferred out on Monday afternoon, the Tuesday tank estimate of 45,000 would actually exceed the accounting expectation of 44,000 before considering mortality. Calling the apparent ten-percent reduction “mortality” would be false. The ledger should carry transfer and sampling events alongside every daily count.
A daily report may also cover several biological stages. If an operator counts “live Z5” on Monday and “live megalopae” on Tuesday, the two categories do not cover exactly the same animals if Z5 larvae persist on Tuesday. If 30,000 of the original Z5 are now megalopae and 15,000 are still Z5, the total observed late-stage population is 45,000, not 30,000. Use stage-specific counts for developmental composition and the union of counted live stages for a broader inventory, ensuring categories do not overlap.
A direct observation of dead bodies might establish minimum observed mortality, but many dead larvae may decompose, be removed in cleaning or be consumed. Thus “dead larvae found: 300” is not necessarily equal to “total mortality: 300.” Separate observed dead, estimated alive, transferred alive and unaccounted difference. This vocabulary prevents an apparently complete numerical report from smuggling in a biological certainty that was never measured.
When an estimate should carry a range
Suppose three independent well-mixed subsamples produce 380, 420 and 460 larvae/L. Their mean is 420 larvae/L. In a 1,000 L working volume, the central inventory estimate is 420,000 larvae. The spread warns us not to quote the full inventory as though every digit were measured. Without a specified statistical design, a transparent descriptive interval based on observed concentrations is simply 380,000 to 460,000, not a formal 95% confidence interval.
If working volume is itself uncertain, the total range widens. Imagine a tank level estimate between 950 and 1,050 L and measured concentration range of 380 to 460 larvae/L. An intentionally conservative combination gives 380 times 950 = 361,000 at one end and 460 times 1,050 = 483,000 at the other. These extremes are not a probability statement. They are scenario bounds under explicitly stated input ranges, valuable for showing how a confident-looking count depends on assumptions.
Readers often want one neat number. Give the central estimate with the method, and keep uncertainty visible nearby. For example: “Estimated 420,000 live larvae from mean concentration 420/L and nominal tank volume 1,000 L; replicate concentrations 380–460/L and volume about 950–1,050 L.” A translator need not invent a precise confidence level to make the interpretation responsible.
How water exchange changes the accounting
A hatchery frequently changes water while the same animals remain in the tank, but screens or sieves may retain larvae imperfectly. If 20% of water is replaced, it is incorrect to deduct 20% of the animals automatically. Such a deduction assumes larvae leave at precisely the same concentration as the water and that there is no barrier or retention system. Equally, assuming zero animal loss requires evidence that the retention setup worked.
The correct sentence is operational: “Twenty percent of tank water was exchanged through a screen; larval loss during exchange was not measured.” If a measured outflow sample establishes an average of 40 larvae/L and 200 L leaves the system, the estimated animal export is 8,000 larvae, with the usual concerns about sampling and flow variation. Now the ledger has an observed basis for the loss estimate rather than an invented one.
Water added changes the concentration even when animal inventory remains stable. Adding 250 L of clean water to a 750 L tank containing 300,000 larvae gives a nominal final concentration of 300 larvae/L, down from 400 larvae/L. A report that calls this 25% mortality would be confusing dilution with death. Always multiply concentration by actual water volume before drawing conclusions about survival.
How to report partial harvests fairly
Megalopae may be removed in several batches as they appear, so “final tank count” can understate the stage’s total recovered output. Suppose 5,000 megalopae are harvested Monday and 4,000 Tuesday, then 3,000 remain in the tank Wednesday. If each harvested batch is unique, the known recovered-or-present inventory across this accounting window is 12,000. But the 3,000 Wednesday estimate could include animals that are not yet megalopae if the sampler combined stages; that would require correction.
The report should state whether harvest numbers were direct counts, estimated from collection volume or inferred from a weight-to-number conversion. If 10,000 megalopae are packed in a bag based on a subsample, “10,000 stocked” is an estimated input, not necessarily a verified direct count. A downstream nursery calculation using this input inherits that uncertainty. If both source and destination independently estimate the same transfer, disagreement between them is a reconciliation question.
For cumulative survival to remain interpretable, do not add the same organisms more than once. A batch transferred from hatch tank H to nursery N is a removal at H and an arrival at N. It belongs in two location ledgers but only once in the cohort-wide inventory. This distinction is familiar in accounting but essential in biology too. Record departures, arrivals, transit mortality where confirmed, and any unexplained difference.
Reconcile equipment, clock and operator changes
A count at 07:00 and another at 19:00 might straddle feeding, siphoning, water exchange, a power interruption or metamorphosis. Even if both rows say “day 12,” they may not describe an unchanged system. The same applies when a new counter, sample device or microscope technique is introduced. A translation should name the event that makes a before-and-after comparison less direct.
Keep local time and, where relevant, time zone in the record; translate AM/PM conventions carefully. If larvae are transferred at 06:30 but the stage count is timestamped 07:00, using the pre-transfer volume or population would be a wrong pairing of measurements. When records contain only dates, mark the order as unknown rather than assuming it from the position of lines in a spreadsheet.
A change of sampler matters as well. One person may identify uncertain objects as late Z5 while another marks them unclassified. A short overlap exercise where both technicians classify the same blinded images can reveal disagreement. Report the agreement measure and category definitions if available; do not say observers “agreed” simply because the daily total counts were similar.
How to compare two hatchery batches without selecting a winner too early
Consider hatchery run A with 300,000 Z1 input and 18,000 megalopae recovered, against run B with 500,000 Z1 input and 20,000 megalopae recovered. Run B has greater absolute output, but A has higher recovery fraction: 18,000/300,000 = 6%, whereas B has 20,000/500,000 = 4%. The word “better” is not a statistic until the goal is specified.
If run A needed twice as much space, energy or labour as B, neither percentage alone answers productivity per unit resource. Conversely, if one run contains larger, developmentally robust megalopae and the other does not, total count may not represent equivalent saleable output. This guide calculates stage counts; a broader operational comparison requires verified quality, cost, timing and facility constraints. A translation should expose that missing frame rather than invent it.
Age and stage also complicate comparison. A day-18 recovery and a day-22 recovery may have different opportunities for further metamorphosis and loss. If two cohorts are at unequal endpoints, label results as interim and final rather than placing them side-by-side as if both have completed the same transition. A sensible comparative statement is conditional: “At their respective reporting dates, these were the observed recoveries; final stage-matched results are unavailable.”
Transfer exercise 1: an apparent improvement caused by dilution
A nursery technician reports 600 larvae/L before topping up an 800 L tank and 450 larvae/L afterwards when the effective volume is 1,100 L. Did the tank lose larvae? Estimate the original inventory as 600 times 800 = 480,000. Estimate the later inventory as 450 times 1,100 = 495,000. The later central estimate is higher by 15,000, or 3.125% relative to the earlier estimate, despite the lower concentration.
The correct interpretation is that concentration decreased but inferred population did not. Given sampling uncertainty, the 15,000 difference does not establish biological population gain; larvae are not reproducing at that stage. The report should investigate sampling variation, the accuracy of the working-volume figures and whether the top-up was measured properly. A disciplined translation says “no demonstrated 25% loss from concentration alone.”
Transfer exercise 2: compute the real nursery recovery
One batch delivers 12,500 unique megalopae to a nursery. At the endpoint, a direct sorting process recovers 8,000 live first-stage crabs and 250 live megalopae, while 150 bodies are seen. What are the ratios? C1 recovery is 8,000/12,500 = 64%. Live stage-inclusive observed recovery is (8,000 + 250)/12,500 = 66%. Confirmed dead individuals recovered equal 150/12,500 = 1.2% of input.
The remaining difference is 12,500 – 8,000 – 250 – 150 = 4,100 individuals, or 32.8% of the input, but that does not identify the causal mechanism. Some animals may have died and not been found; others may be hidden or lost in the counting method. The wording “unaccounted at the endpoint” is more accurate than “32.8% killed by cannibalism,” even when cannibalism is a plausible and important biological risk.
Transfer exercise 3: a stage mix is not a cohort yield
A microscope sample has 80 megalopae, 100 Z5 and 20 Z4, all classified live. Calculate percentages: 80/200 = 40% megalopa; 100/200 = 50% Z5; 20/200 = 10% Z4. These are sample composition percentages with denominator 200 observed live larvae. If the starting Z1 cohort was estimated at 40,000, do not write “40% of hatch survived to megalopa.”
Suppose, separately, there is a comparable tank inventory estimate of 1,600 live megalopae on that date. Then the apparent megalopa inventory relative to the starting cohort is 1,600/40,000 = 4%, still a snapshot rather than a cumulative developmental success rate. The difference between 40% and 4% comes from the denominator, not a calculation mistake. This is one of the most valuable distinctions to preserve in a multilingual report.
Transfer exercise 4: why pooled recovery is weighted
Nursery A starts with 2,000 megalopae and yields 1,600 first crabs. Nursery B starts with 8,000 and yields 4,000 first crabs. Their individual recoveries are 80% and 50%. The simple mean of 65% treats each nursery as equally important despite a fourfold difference in stocked animals. Pooled recovery equals (1,600+4,000)/(2,000+8,000) = 56%.
The correct number depends on the question. “Average nursery percentage” may be an unweighted mean when each nursery unit is the statistical subject, but “fraction of all stocked megalopae recovered” needs weighting by input. A report should not silently switch between those interpretations. The difference can change managerial conclusions and must survive translation.
Transfer exercise 5: reconstruct a chain and check it
Suppose a hatch has estimated 240,000 Z1 larvae. It yields 24,000 unique megalopae, of which 18,000 are stocked in a nursery and 6,000 are sold or sent elsewhere. From the 18,000 stocked, 9,900 C1 are recovered. Z1-to-megalopa recovery is 24,000/240,000 = 10%; nursery conversion is 9,900/18,000 = 55%; and retained-chain C1 output relative to original Z1 is 9,900/240,000 = 4.125%.
Multiplying 10% by 55% would give 5.5%, which is wrong for the retained C1 chain because not all 24,000 megalopae entered the nursery. The missing fraction is the allocation 18,000/24,000 = 75%. The check is 10% times 75% times 55% = 4.125%. This exercise shows why transfers and sales belong in the biological accounting chain, even if all numbers are correctly typed into a spreadsheet.
Transfer exercise 6: report uncertainty without manufacturing precision
The operator has four estimated stage counts: 420, 390, 440 and 410 larvae/L, plus a nominal volume of 900 L. Their mean concentration is 415 larvae/L and nominal inventory is 373,500 larvae. Report the four observed results rather than just the mean. Their sample range is 390–440 larvae/L; at 900 L this corresponds to 351,000–396,000 as an illustrative range of inventory estimates.
That observed spread is not a verified confidence interval for tank abundance. If the sampling method systematically misses larvae near the bottom, all four may be biased. To make a formal uncertainty statement, define the sampling design and appropriate estimator, including measurement and spatial heterogeneity. A translation can remain useful and accurate without inventing a statistical certification.
Frequently asked questions about stage counts
Does Z5 mean the fifth day of culture? No. Z5 identifies a zoeal developmental stage, although in some systems a day-like code may be used elsewhere. Read the local legend. A fifth-day observation and fifth zoeal stage are not interchangeable, and confusing them can transform a chronology into an anatomical classification. The safest translated text states both stage and elapsed days separately if both are known.
Is a megalopa a small adult crab? Not yet. Megalopa is a transitional larval stage before the first juvenile crab instar, with different structures and habitat use. A label that calls all megalopae “juveniles” may cause the reader to assume they are already at the nursery endpoint. Use “megalopa” with a brief explanation and reserve C1 or first crab for the next stage.
Can we add all daily megalopa counts to get production? Only when the numbers are unique recovered batches, not repeated tank inventories. Daily snapshots can count the same individual repeatedly. A cumulative harvest record should give the number removed at each step and identify the receiving container or unit so an individual isn’t counted again during transfer.
Does a high sample proportion mean good survival? Not necessarily. A sample that is 90% megalopae could come from a tank with only a small surviving population. It describes a stage mix at one moment. To discuss recovery, compare appropriately estimated unique outputs against a declared original input or stage-specific starting population.
Is a dead count enough to explain the missing inventory? Usually not. Dead animals may disappear through decomposition, predation or cleaning and may not be recovered by the counting method. Conversely, animals can survive but hide or be undercounted. Report observed dead counts and unaccounted inventory separately, and do not assign a cause without evidence.
Can two reports use “survival” but mean different things? Yes. One may measure Z1-to-megalopa recovery, another megalopa-to-C1 recovery, and another the final output divided by starting Z1. Translate the complete numerator and denominator next to each percentage. A label such as “55% survival” without the stage interval and endpoint is incomplete.
What should a reader do when a number has no unit? Ask for the source record, sampling volume, concentration factor and actual tank volume if relevant. Do not infer a unit from the apparent size of the number or from another facility’s convention. Mark it as “unit not stated” until resolved and do not use it in a pooled calculation.
What if a stage label seems biologically impossible for the date? Check species identification, local coding, time since actual hatch and whether records from separate cohorts were merged. An implausible entry is a validation prompt, not automatic proof of fraud or operator error. Preserve the original label in the audit trail while investigating.
A final reliability check before sharing a translated report
Before the report goes to a buyer, supervisor, student or collaborating hatchery, read every percentage aloud with its full denominator. “Nine thousand first crabs out of eighteen thousand megalopae stocked” is self-explanatory; “fifty-percent success” is not. Then check that total units balance: litres with larvae per litre, organisms with organisms, counts with stage categories, and calendar time with the proper culture event.
Verify that each transfer has one source and one destination, and that repeated samples were not added as unique production. The resulting translation should preserve the difference between direct observations, estimated inventories, stage-specific ratios, and biological explanations. Those distinctions are the practical technology of reliable technical language: they let a reader reconstruct what happened without being forced to trust a headline percentage.
Sources and further reading
FAO, Mud crab aquaculture: a practical manual (Shelley and Lovatelli, 2011). Guidance on life stages, hatchery monitoring, rearing and nursery management.
SEAFDEC, Biology and hatchery of mud crabs Scylla spp.. Specialist stage and hatchery reference.
SEAFDEC, Seed production of mud crab Scylla serrata juveniles. Reported cohort transitions and mortality challenges in a documented production study.
eduKateSG, How to Translate Easily series index. Navigate other technical reading guides.
eduKateSG, How English Works: Vocabulary and Meaning. Further reading about precise word meaning in context.
eduKateSG, How to Read Shrimp Postlarval Volumetric Counts. Related discussion of subsampling and numerical extrapolation in another crustacean.