From landing forms to defensible catch and trade interpretations
Seahorse fishery records need to be read as observations about fishing activity and retained or observed catch, not as direct counts of all seahorses living in the sea. A landing survey may record a boat trip, gear, time fished, locations, number of seahorses sorted from catches, body measurements and interview responses. A separate trade record may list specimens by product type, species and lawful documentation. Those records answer different questions and cannot be merged without tracing the animals and the quantities involved.
The essential interpretive step is catch per unit effort, or CPUE. It needs a consistent count and a denominator that measures comparable fishing effort. Changes in gear efficiency, effort, fishing ground, observation method or reporting can change CPUE even when underlying seahorse abundance has not changed. The iSeahorse Landings Trends Toolkit from Project Seahorse explicitly warns that landings alone are insufficient and that the resulting indices must not simply be declared true population sizes. The guide below uses invented teaching cases only; it neither documents an actual fishery nor advises wildlife trade.
Quick summary: identify location and survey frame; record trip and vessel IDs; distinguish trips, tows, hours, boat-days and gear-specific effort; count confirmed seahorses separately from unknown specimens and other bycatch; preserve zero-catch trips and missing interviews; define landed, discarded, kept, sold and exported; calculate CPUE with an explicit effort unit; separate species confidence, size and sex classification; check selection bias, observation gaps, market conversion and chain-of-custody; read CITES paperwork as compliance evidence subject to applicable national rules, not proof of biological sustainability.
A field-by-field reader’s guide
Landings-site identity
A port, landing beach, fish market and fishing ground are different locations. A survey station describes where the observation happened, not necessarily where animals were caught. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Preserve site name, coordinates where appropriate, survey date and responsible sampling team. A market may aggregate catch from a wide geographical area. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A record of 15 seahorses observed at one town’s market cannot prove all 15 were captured in the waters immediately beside that town. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Site selection
Ports may be sampled because seahorses were previously observed, because they are accessible or because they represent a broader fishery. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Selection affects inference. Avoid presenting purposive monitoring as a random sample of every fishing port in a region. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Three ports chosen for known high seahorse trade can track those ports but cannot automatically produce an unbiased national average. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Trip identifier
The same fisher and vessel can undertake many trips. A trip should have a defined departure, return and unique record. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Two interviews about one trip are not two independent fishing events. Cross-reference date, vessel and trip codes to avoid duplication. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Vessel A’s Monday landing appearing on two forms represents one trip even if two enumerators recorded it. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Fisher interview
Self-reported catch and observed sorted catch provide different kinds of evidence. Memory, incentives and incomplete recall can affect interviews. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Keep reported values distinguishable from physically examined specimens; do not silently upgrade remembered counts into directly verified measurements. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A fisher saying ‘about 20 last week’ provides an estimate over an imprecise window, not 20 separately counted seahorses in the current landing. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Vessel category
Vessel size, operating range and capacity can influence trip duration and fishing grounds, so all boats may not be comparable. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
A coarse label such as ‘small boat’ should not be turned into a precise engine-power class without documentation. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
An apparent increase in bycatch after larger vessels join the sample may partly reflect different effort and locations. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Fishing gear
Bottom trawls, seines, pots and other gear differ in catchability and exposure to seahorses. Gear descriptions must be kept exact. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Do not pool raw catch-per-hour numbers across gear types as though each hour has the same biological meaning. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
One seahorse per trawl-hour and one per pot soak-hour use denominators that cannot automatically be compared. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Target catch
Seahorses may be the target of a fishery or incidental bycatch from fishing for another species. This distinction matters for interpretation. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Use terms ‘targeted’, ‘incidentally captured’ and ‘unknown’ according to evidence, not assumption from a merchant’s product description. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A trawler fishing primarily for shrimp may report seahorses sold later; sale does not make them the trip’s intended target. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Tow count
A tow or haul is an effort unit only when its start, end and gear type are defined consistently. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Three short tows are not necessarily equivalent to three long tows. Preserve both counts and duration where possible. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A trip with eight tows at 30 minutes each has four recorded tow-hours, assuming no interrupted operations. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Tow duration
Time actively fishing differs from transit, setting and recovery. The start and end points for effort require a rule. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Report the method of timing. Comparing elapsed vessel hours with active gear hours can distort CPUE. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A six-hour trip with three hours towing and three hours transit supplies three trawl-hours, not six, under an active-towing denominator. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Soak effort
For passive gear, gear count multiplied by hours deployed is often a useful measure, but missing or damaged gear complicates it. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Translate pot-hours or net-hours carefully and state whether incomplete deployments were excluded. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Ten traps soaking eight hours each represent 80 trap-hours under a simple exposure definition; they do not equal eight ordinary trip-hours. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Effort standardisation
Effort must have consistent units before catches are compared through time or between locations. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
If vessel or gear efficiencies differ greatly, simple division may still require additional statistical adjustment. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A two-seahorse trip taking one hour is 2/trawl-hour; a six-seahorse trip taking six hours is 1/trawl-hour, reversing the raw-count ranking. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
CPUE numerator
The numerator may mean all seahorses observed, retained specimens, landed specimens or a species-specific subset. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
State which group was counted. A count that excludes animals discarded before landing cannot be represented as total capture. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A trip catches eight seahorses, discards three and lands five; landed CPUE based on five is not total capture CPUE based on eight. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
CPUE denominator
Trips, fishing hours, tows and gear-days are different denominators. They cannot be interchanged without supporting data. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Give the unit alongside every CPUE value, and keep the definition stable between years. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Forty observed seahorses from 20 trawl-hours equals 2 per trawl-hour, not 2 per boat-day or 2 per tow unless those denominators also happen to equal 20. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Zero-catch trips
Trips with zero observed seahorses are informative when observation coverage was adequate; excluding them inflates average CPUE. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Retain zeros while distinguishing cancelled or unsampled trips. A spreadsheet blank must not be silently translated as zero. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
In ten sampled trips with total catch 30, omitting five true zero trips makes the average per trip 6 rather than the correct 3. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Unobserved trips
A fisher may operate on days without interviews. Those missing events can differ from observed trips. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Report number of eligible trips, number sampled and the reasons trips were missed. Assumptions about missingness affect uncertainty. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
If 20 of 30 eligible trips were interviewed, a 20-trip mean is not automatically a complete 30-trip census. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Unsold bycatch
Seahorses can be discarded, retained for personal use, mixed into low-value catch, sold fresh or dried later. A market record need not capture them all. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Translate ‘landed’ and ‘sold’ as distinct process stages. Do not infer total catches from sale volumes without loss and selection accounting. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A dealer purchasing ten dried specimens cannot, by that fact alone, establish how many live seahorses were originally caught. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Species count
A catch may include several Hippocampus species and unidentified animals. Grouping them hides species-specific trends. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Preserve confirmed IDs and unresolved specimens with a confidence scale. A generic category should not be assigned the most common local species by default. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Nine confirmed species A, four confirmed species B and three unknown yield 16 total observed seahorses but only nine verified species-A individuals. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Photographic identification
Clear images with diagnostic characters are more useful than colour descriptions alone. Dried specimens can be especially difficult to identify. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Record image ID, viewed characters, expert determination and any subsequent corrections. Do not let copied labels become independent confirmations. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Two spreadsheets bearing the same mistaken field identification do not constitute two independent taxonomic validations. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Individual size measurement
Seahorse height or torso length may use specialised measurement landmarks; curved tails and posture make naive ruler measurements unreliable. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Retain the original measurement definition and units. The CITES-related 10-cm discussion must not be misrepresented as a universal legal size limit. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
If 12 cm is total height and 8 cm is torso length, they are not two competing measurements of the same anatomical dimension. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Minimum-size interpretation
Management guidance has historically discussed a 10-cm minimum height for many seahorses, but species, national implementation and trade context vary. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Read the actual instrument and present-day local rule; do not turn a scientific recommendation into a globally binding harvest law. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A specimen of 9.5 cm is a measurement, not by itself a legal determination without checking the applicable law and definition. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Sex ratio
A sample may contain identifiable males, females and animals of undetermined sex, including juveniles. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Report the denominator and exclude uncertain categories only when the statistic is explicitly a sex ratio among sexed individuals. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
With eight known males, six known females and six unknown, males are 8/14 among sexed animals but 8/20 among the full observed sample. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Visible pregnancy
A swollen male brood pouch can indicate reproductive condition but may be recorded without a diagnostic confirmation. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Translate ‘apparently pregnant’ honestly. The proportion of apparently pregnant males in landings depends on catch selectivity and timing. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Four pouch-swollen males among ten males in a sample is a 40-percent observed appearance fraction, not proof of a population-wide pregnancy rate. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Mean body size
A mean calculated from photographed and measurable animals may omit damaged, obscured or unusually small specimens. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
State sample size and whether each length was independently measured. Survivorship and market selection can shape the observed size distribution. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A mean of 11 cm among eight measurable specimens does not prove the same mean for the six specimens with no scale record. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Size classes
Histograms require mutually exclusive category boundaries and consistent rounding. Exactly-on-boundary specimens need a defined rule. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Do not change open and closed interval notation during translation; doing so can alter a frequency table. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A fish measured exactly 10.0 cm must be assigned consistently to either the lower or upper bin under the documented convention. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Count versus mass
Trade documents may use pieces, kilograms, dried weight, wet weight or batches. Unit changes require validated conversion factors. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Do not estimate pieces from kilograms using an assumed average specimen mass unless the range and uncertainty are reported. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A 2-kg dried shipment cannot be declared to contain 200 specimens merely because one convenient estimate assumes 10 g each. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Source codes
CITES and domestic documents may distinguish wild, captive-bred or other source classifications, but a code requires context and official definitions. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Avoid interpreting a source code as independently verified genealogy or animal welfare history. Check the current official coding rules. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A document marked captive origin is a declared classification that still depends on the evidence and applicable regulatory assessment. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Appendix status
Seahorses of the genus Hippocampus are listed under CITES Appendix II, making internationally traded specimens subject to treaty procedures and national implementation. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Translate appendix listing as a regulatory framework rather than an automatic ban on all trade. Check the current official species listing and country-specific controls. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A shipment bearing a permit may meet one documentation requirement, but its legal status cannot be concluded from a photograph of an unverified document alone. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Export and re-export
An export from the country of origin differs from a re-export after import through another country. Trace the chain before inferring provenance. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Keep origin country, exporting country, re-exporting country and dates separate. The port of departure is not necessarily the biological origin. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A shipment leaving country B may contain specimens originally collected in country A; ‘exported from B’ does not settle origin. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Permit validity
A permit identifier, issue date, validity period, quantity and listed taxon all need verification against the actual official documentation. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Treat a permit entry as evidence to inspect, not proof that every presented specimen belongs to the declared shipment. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Two consignments showing the same permit number warrant cross-checking; a duplicated number alone does not establish a lawful or unlawful outcome. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Chain of custody
Collector, landing station, wholesaler, processor and exporter may each report quantities in different units and stages. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Reconcile transfer receipts, lot labels and processing losses without double counting sales at successive steps. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
If 100 dried seahorses move from trader A to B then to C, the three invoices describe one lot’s movements, not 300 newly caught animals. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Catch trends
A decline in observed CPUE can signal concern but can also reflect gear changes, fishing-ground shifts, weather or reporting practice. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Use repeat sampling with similar protocols and covariates. Avoid a causal or population-trend declaration based solely on raw annual totals. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
Twenty seahorses recorded in one month and ten the next may represent less fishing effort rather than an actual halving of abundance. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Observer and reporting bias
Fishers, observers and markets may report different subsets for economic, regulatory or logistical reasons. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Triangulate interviews and physically checked samples where lawful and ethical. Explain discrepancies without assuming bad faith or perfect accuracy. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
An enumerator who arrives only after peak landings may miss early arrivals even when every observed interview is accurate. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Sampling uncertainty
Observed rates from small or unevenly sampled trips can have high variance and strong clustering within vessels or days. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Report trip counts, variation, species-specific totals and intervals where justified. Avoid unwarranted decimal precision. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A mean of 1.8 seahorses per tow from only five tows should not be presented as a stable regional biological rate. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Conservation claims
A landing dataset can identify useful warning signals, but a claim about population decline or success of management needs broader evidence. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
Relate catch data to effort, habitat, survey coverage and independent monitoring. Avoid using trade paperwork as a substitute for ecological assessment. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A low reported export total after stricter enforcement could reflect reduced trade, displaced routes or reporting changes; it is not alone a measure of wild recovery. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Respectful interviewing
Surveyors should document informed participation, protect sensitive identities and follow local legal and scientific guidance. The first translation test is whether the statement describes what was seen, what someone reported, what a calculation assumes, or what a regulator has formally established. Those evidentiary categories are different. Keep the field’s original term, unit, date, source and confidence in a transparent explanatory note so that a later analyst does not unknowingly convert a provisional entry into a stronger claim.
An evidence-based report needs transparent methods without exposing individuals or encouraging unlawful collecting. Where two surveys or commercial records are compared, align geography, species category, effort, transaction stage and reporting conditions before computing a trend. The same count can have a different meaning when the sampling or commercial pathway changes. Record missingness rather than replacing blanks with guessed zeros, and keep uncertainty visible next to any numerical estimate or legal interpretation.
A translated dataset may retain anonymous vessel codes while removing names and exact sensitive locations under an approved data-governance plan. The useful habit is to reconstruct the data trail: identify its origin, check for repeated counting, write the numerator and denominator, recompute any stated rate, and ask what the figures cannot observe. When a sentence sounds more certain than the underlying form, narrow the English claim instead of smoothing away the qualification. This is how a readable translation remains faithful to evidence.
Checked examples, independent practice and questions
Fully checked worked example: A landing monitoring team interviews eight qualifying trawl trips. Recorded active trawl-hours are 3, 4, 5, 4, 2, 6, 3 and 5, totaling 32 trawl-hours. Seahorses physically recorded for those trips are 0, 3, 4, 2, 0, 5, 1 and 1, totaling 16. Pooled landed CPUE is 16/32 = 0.5 observed seahorses per trawl-hour. This is not automatically equal to the arithmetic average of the eight individual trip CPUEs, because their hours differ; nor is it an estimate of every seahorse captured, because discards and unsorted bycatch may be missing. Two zero-catch trips belong in the denominator. If a later survey uses boat-days, it must not be compared without a suitable effort-standardisation procedure.
Independent transfer problem one: A sample has 25 seahorses observed from ten trawl-hours in period A and 30 from twenty trawl-hours in period B. The raw catch increases, but how does CPUE change? Answer: period A is 25/10 = 2.5 seahorses per trawl-hour. Period B is 30/20 = 1.5. The observed rate falls by 1.0 per trawl-hour, or 40 percent relative to A. This is an index difference, not proof that the population declined 40 percent.
Independent transfer problem two: A fisher reports catching 14 specimens: nine landed, three discarded alive and two retained for personal use. A market register contains only nine. What is the correct numerator for market-landed CPUE versus reported total-capture CPUE? Answer: nine for the observed market-landed quantity, and 14 for the self-reported total capture if the interview is credible and refers to the same defined trip. Keep their evidence type and coverage distinct; the market register alone cannot validate the five absent animals.
Independent transfer problem three: A survey records 18 confirmed H. comes, six confirmed other Hippocampus and four unidentified animals. A draft translates this as 28 H. comes. What repair is needed? Answer: the complete seahorse total is 28, but verified H. comes is 18, verified other species six and unidentified four. Reassigning the unknown four and the other six changes species-specific evidence without justification.
FAQ: Does CPUE equal true population abundance? No, it is at best an index under assumptions about effort and catchability. Is every seahorse at a fish market necessarily local? No. Can all exported pieces be added across wholesalers as separate catch? No, chain-of-custody movements may represent the same items. Is Appendix II equivalent to a universal trade ban? No; applicable permits and national restrictions must be checked. Does a 10-cm figure have the same legal force everywhere? No; verify the actual current instrument.
Primary methodological sources
Sources and adjacent learning: The Project Seahorse Landings Trends Toolkit, https://projectseahorse.org/wp-content/uploads/2022/08/iSeahorse_Landings_Manual_English_LowRes_1.0.pdf , is the primary method reference for effort and landing surveys, and https://projectseahorse.org/ provides conservation context. CITES provides its seahorse identification guide at https://cites.org/eng/virtual-college/guide-identification-seahorses and an Animals Committee scientific discussion of minimum-height recommendations at https://cites.org/sites/default/files/eng/com/ac/19/E19-16-2.pdf . These documents are not substitutes for current national law. For a complementary underwater observation guide that measures habitat rather than fishing effort, see https://edukatesg.com/2026/11/01/how-to-translate-easily-seahorse-habitat-transects-holdfast-detection-density/ .