Settlement is an event, not the end of the story
Sea urchin settlement is a transition in the life cycle: a competent larva encounters a suitable cue or surface, leaves the swimming phase and begins the metamorphosis that leads to a benthic juvenile. A count of larvae seen resting near a plate is not necessarily a count of successfully metamorphosed juveniles. A high initial settlement fraction is not the same as high survival several weeks later. A technical translation must keep initial competence, attachment or settlement, completed metamorphosis and established juvenile retention separate.
This guide explains how to read sea urchin settlement-plate reports, biofilm-cue trials, percentage settled, spatial quadrat surveys, metamorphosis scoring and juvenile nursery survival. It is written for readers of scientific methods, hatchery dashboards and cross-language technical summaries. All numbers in the worked examples are invented teaching cases. Published research on Tripneustes gratilla shows that very strong laboratory settlement cues do not necessarily produce better post-settlement survival, making careful endpoint language essential.
Quick reading sequence: find the number of competent larvae introduced; define what the study calls ‘settled’; identify the census time and surfaces sampled; convert sampled quadrat counts into defensible estimates; distinguish initial attachment from a morphological juvenile endpoint; calculate survival using the correct earlier population; account for plate transfers and incomplete detection; and state whether a result describes a laboratory cue trial, a nursery tank or a field release.
The biological boundaries around metamorphosis
Before settlement, a typical sea urchin pluteus larva swims and feeds in the plankton. As competence develops it can respond to environmental cues associated with future bottom habitat. Settlement begins the shift into a substrate-associated organism, and metamorphosis involves substantial changes in body form. A newly formed juvenile then needs food, shelter and suitable environmental conditions. The exact timing and behaviour vary by species and setup.
In casual speech, larvae might be described as ‘settled’ when they stop swimming, but a scientific paper may define settlement more strictly by post-attachment morphology. Another report may use the phrase ‘metamorphosed’ only after specified juvenile structures develop. These words cannot be translated as identical terms without losing the operational definition of the assay.
Some observations classify larvae as competent before exposing them to a settlement cue. The competence percentage may be estimated from a sample, not every animal in the culture. If 30% of a culture is competent, the relevant cue-exposure denominator might be all introduced larvae or only the subset selected as competent. The article or contract must say which.
Biofilms and why the cue matters
Sea urchin larvae can respond to surfaces bearing biological films, including algae and associated microorganisms, but results are species- and context-specific. Research by Dworjanyn and Pirozzi on Tripneustes gratilla examined responses to macroalgae, marine biofilms and conspecific signals. It found evidence that bacterial components on algal surfaces can matter for induction in that species and those experimental conditions. This is not a universal recipe for creating a hatchery plate.
A plate may be called ‘conditioned’ because it has been exposed to a defined culture environment. That label alone does not establish which organism or chemical cue is present, nor how well it will support post-settlement feeding. A record should identify conditioning method, substrate, treatment duration and the criterion for a successfully induced animal if those details are relevant to interpreting the result.
The phrase ‘biofilm treatment increased settlement’ may mean the percentage responding in a controlled assay increased. It does not necessarily mean more animals survived in the nursery or that the cue is approved for deployment in a natural environment. These later steps require their own evidence and authorizations.
The experimental unit is often the tank, not the plate count
Imagine five settlement plates inside one tank with a treated biofilm and five plates inside another tank with untreated surfaces. There are ten plates, but the experimental design may still contain only two independent tanks, one per treatment. Plates share water, environmental conditions and the history of their tank. Treating all plates as independent biological replicates may produce unjustified statistical confidence.
A translation should distinguish plate subsamples from independently randomized replicate tanks. If a study has four tanks per treatment and measures five plates in each, it may summarize each tank first and then compare the four tank-level observations. This preserves the design rather than pretending that twenty plates are twenty independent treatment environments.
Similarly, repeated photography of the same plate over several days provides longitudinal observations of one surface, not fresh replicates each day. The source’s experimental-unit terminology matters more than the number of rows in a spreadsheet.
The correct settlement fraction
The most direct settlement fraction for a defined exposure test is the number scored as settled divided by the number of larvae introduced into the test, under the author’s explicit inclusion and exclusion rules. If 1,200 competent larvae are introduced and 420 meet the defined settlement criterion at the prescribed census, observed settlement is 420/1,200×100 or 35%. That is not a 35% final juvenile yield.
If only 900 of the 1,200 larvae were competent but the other 300 were nevertheless introduced, the proportion settled among all introduced remains 35%. If the researcher instead wants settlement conditional on competence and has actually traced which animals were competent, the numerator and denominator may be different. Do not substitute the assumed competence count for the real exposure denominator without seeing the design.
Some larvae may be unobserved because they remain swimming, settle on another surface, hide, or become damaged. The difference between introduced and visibly settled is not automatically a mortality count. A valid interpretation depends on where the observers looked and how the study categorised unaccounted individuals.
What counts on a plate cannot count elsewhere
In a settlement tub, newly settled urchins may occupy plates, side walls, the bottom and equipment surfaces. A count restricted to plate faces does not necessarily represent the total settled population. A report may deliberately measure settlement preference for plates, or it may estimate settlement throughout the whole tub using samples from different surfaces. Those are different outcomes.
Suppose plate faces contain 80 observed juveniles, while a complete wall census finds 25 and a defensible bottom estimate finds 45. If all categories are mutually exclusive and covered, the estimated total is 150. ‘Plate settlement’ would be 80, or 53.33% of the total observed or estimated settled animals. A statement that only 80 juveniles settled in the vessel would be incorrect if the study also counted the other surfaces.
Careful translation distinguishes plate count, bottom-quadrat estimate, side-wall census and full-container estimate. A reference in a caption to ‘total per replicate’ should not be replaced with ‘total per plate’. The number changes its spatial boundary.
Sampling a quadrat without assuming perfect detection
Researchers may count juvenile urchins within a known small area and extrapolate to a larger surface. If a 10 cm by 10 cm quadrat holds 12 visible juveniles, its area is 100 cm². For a uniform, fully observable surface of 1,000 cm², a simple area-ratio estimate would be 12×(1,000/100), or 120 juveniles. This is only a teaching calculation: distribution can be patchy, detection imperfect and quadrats nonrandom.
To produce a stronger estimate, a study should state how many quadrats were sampled, whether their positions were random or systematic, the area represented, and what happens where juveniles hide in biofilm or around structures. If observers use haphazard quadrat placement, the resulting extrapolation has different inference limits from a validated probability sample.
Unit conversion is another hazard. An area of 50 mm by 50 mm is 2,500 mm², or 25 cm², not 2,500 cm². A translated figure caption that drops the square sign changes the spatial arithmetic completely. Always preserve cm², mm² and the measured surface geometry.
Settlement, metamorphosis and short-term survival as a chain
Consider a hypothetical vessel stocked with 10,000 larvae. At the first defined census 4,000 qualify as settled. Later 2,800 are scored as morphologically metamorphosed juveniles, and 1,400 are alive at a further nursery checkpoint. The observed settlement percentage relative to introduced larvae is 40%. The conditional fraction reaching the later metamorphosis endpoint among settled is 2,800/4,000, or 70%, assuming linked cohorts.
Nursery survival among the scored metamorphosed juveniles is 1,400/2,800, or 50%. Cumulative proportion remaining at the nursery checkpoint is 1,400/10,000, or 14%. Multiplying 0.40×0.70×0.50 reproduces 0.14. The intermediate percentages cannot be added, and the endpoint remains short-term nursery retention rather than lifetime survival or ecological establishment.
In real records each stage may be estimated from separate sampling procedures. In that case the multiplication is an accounting model whose accuracy depends on matching cohorts, timing and definitions. Do not present its resulting whole-tank count as individually marked animals unless individual tracking actually took place.
Why a very strong cue can still fail commercially
A peer-reviewed Southern Cross University study on Tripneustes gratilla compared settlement stimuli and examined post-settlement survival. The study showed that some cues could greatly increase the proportion of larvae responding at settlement, yet high mortality after settlement could erase the apparent advantage over longer follow-up. That result demonstrates an important principle: optimizing one early endpoint does not guarantee optimization of the final nursery output.
A translation that reports only ‘over 90% settlement’ from an experimental treatment while omitting follow-up mortality can give a distorted impression of success. The time and context matter: a controlled laboratory assay may have a small number of larvae and a narrow observation window, while production involves larger tanks, changing films, food availability, maintenance and repeated handling.
Describe the outcome as ‘high induction of settlement under the tested conditions’ if that is what was observed. Reserve ‘superior juvenile production’ for a measured increase in surviving juveniles at a defined final checkpoint. The difference may decide whether a reader understands the study at all.
Plate area, carrying space and juvenile density
Once juveniles have settled, a nursery may describe their density as juveniles per square centimetre or per square metre of available substrate. These are area densities, not water-column densities. If 300 juveniles occupy 0.6 m² of defined usable surface, the areal density is 500 juveniles/m². Converting that to 0.05 juveniles/cm² is possible because 1 m² contains 10,000 cm².
A report may give the number of physical plates without specifying their active coated area. Ten plates of different sizes do not represent a consistent settlement area. A rectangular plate with both sides accessible may have roughly twice the geometric face area of a one-sided plate, but brackets, edges, fouling and contact surfaces complicate the actual usable area. State the surface definition.
Density alone cannot establish crowding injury, food supply or predicted growth. It is a descriptive measure that needs species, stage, substrate conditions and observation date. Do not translate a geometric surface estimate as biologically available grazing habitat unless the study validated that interpretation.
A transfer between settlement plates and a nursery tank
A hatchery report may move settlement plates with attached juveniles into another vessel. Later counts could describe retention on the original plates, survival of the entire moved cohort, or occupancy of the destination tank. Some animals detach and reattach elsewhere; others may be hidden in biofilm. If a plate shows fewer animals after transfer, loss from that surface is not automatically mortality.
Imagine 1,000 juveniles initially observed on twenty plates. After transfer, 720 remain visible on those plates, 110 are observed on the new tank walls and 50 on its bottom. The 880 observed individuals are the known visible remainder within the new vessel. The 120 difference is unaccounted, not all proven dead. Reporting ‘72% survival’ from plate-only counts would be misleading if the intended census covers the entire new tank.
Where detailed animal detection is impossible, label the result an apparent count or recovery estimate and describe surveyed surfaces. A robust nursery ledger includes movements, areas examined, time of census and detection limitations.
Feed transition after settlement is a separate question
Early benthic juveniles can graze on suitable surface-associated food, including biofilms, and may later transition to larger algal feed according to species and size. NOAA and Hawaiʻi’s Division of Aquatic Resources describe this developmental transition in their collector urchin rearing programme. It is relevant context for why the settlement surface may serve as both cue and initial feeding habitat.
The fact that a surface induces settlement does not prove that it supports sustained growth. A treatment could attract many larvae but supply inadequate suitable food or alter later conditions. Growth and survival must therefore be followed beyond the first settlement census. The translation should not promise successful feeding from the word ‘biofilm’ alone.
A field-facing report may include animal size at transfer, size at release and time in the nursery. Those describe additional stages after settlement and should be kept outside the settlement-rate numerator. A scientifically coherent article can connect the sequence without collapsing its distinct endpoints.
Reading size distributions instead of one mean
Juvenile urchin size may be measured by test diameter, sometimes with or without spines according to a stated convention. That measurement differs from larval arm length and from a simple visual description of the animal. If a size distribution is skewed, the mean can hide a large number of small animals and a smaller number of rapidly growing ones.
Suppose five juveniles have measured diameters of 1, 1, 2, 2 and 9 mm. Their arithmetic mean is 3 mm, but four of five animals are no larger than 2 mm. A purchasing or release rule that depends on a minimum size should not use the mean alone. Report counts or proportions meeting the criterion along with the measurement convention.
If the threshold is at least 2 mm, three of the five hypothetical animals qualify, or 60%. That percentage is a number fraction among the observed sample, not a mass fraction or the fraction of the entire hatchery’s stock. Preserve sample selection and date in the English translation.
Nursery retention needs a clear observation interval
A statement such as ‘90% juvenile survival’ should include its time interval and starting stage. Survival over the first two days after settlement does not establish survival over a month. Under a simplifying constant daily survival assumption, 90% surviving a two-day period would not justify simply carrying 90% forward to day thirty. That would require repeated observations and a defensible time model.
A more direct teaching example starts with 2,000 observed metamorphosed juveniles. After seven days 1,600 remain within a fully documented census, giving 80% retention. At day twenty-one 1,000 remain, giving 50% cumulative retention from the first count and 62.5% survival over the interval from day seven to day twenty-one. These three figures must be labelled to prevent misunderstanding.
Without tracking losses or detection completeness, call them apparent or observed retention rather than absolute biological survival. Underlying conditions and counting methods can change between censuses. Keep the uncertainty at the stage where it arises.
From hatchery release to ecology: no shortcut
Sea urchins are deployed in some conservation programmes to graze invasive algae, including a state-operated programme in Hawaiʻi. State material reports major cumulative releases of hatchery-raised collector urchins. A number released is a verified operational output when supported by the programme’s inventory, but it is not the same as the number surviving one year later or the number reproducing in the wild.
A release ledger should contain release date, species, size, place, quantity, handling conditions and monitoring design. Follow-up surveys can examine persistence and environmental effects, but an observed reduction in invasive algae is not automatically attributable to a particular hatchery batch without suitable causal evidence. Multiple environmental and management processes may operate together.
Write ‘released’ for a documented outplanting event, ‘observed alive at survey’ for a sighting, and ‘estimated survival’ only when the survey and detection model justify the estimate. The words sound close in casual reporting but encode very different scientific claims.
Transfer problem: two settlement cues with unequal starting groups
A trial assigns 600 competent larvae to treatment A and 1,200 to treatment B. At the first census A has 360 animals meeting the settlement definition, and B has 600. A’s observed settlement is 360/600×100, or 60%; B’s is 600/1,200×100, or 50%. B produces more settled animals in absolute numbers because it began with twice as many larvae, but A has the higher percentage response. These facts are consistent.
If a summary says ‘B outperformed A because 600 settled rather than 360’, it may be describing output volume but not the experimental response fraction. If both treatments later show only 50% survival of their settled juveniles, A ends with 180 and B with 300. The more useful endpoint depends on the question: per-larva treatment effectiveness, output per tank, or cost per live juvenile. The translation must state the choice.
Transfer problem: count both faces of an inclined plate
A rectangular settling plate measures 20 cm by 30 cm, giving 600 cm² per face. If both faces are accessible, nominal face area is 1,200 cm², not counting edges. Suppose 240 juveniles are visible on the two faces combined. Nominal observed density is 240/1,200, or 0.20 juveniles/cm². Reporting 240/600=0.40 assumes only one face and doubles the estimate.
The precise usable area could differ if part of each plate is submerged in sediment, pressed against a frame or covered by inaccessible biofilm. Identify whether the original report used projected area, both sides, only the inspected side or a more detailed surface estimate. The operator’s convention belongs next to the density result, especially where plates of differing shape are compared.
Transfer problem: five quadrats and one unusually dense patch
Five equal 25 cm² quadrats contain 4, 6, 5, 24 and 6 visible juveniles. Together they hold 45 in 125 cm², giving a sampled mean density of 0.36 juveniles/cm². If the surveyed surface area is 1,000 cm² and the sampling design supports extrapolation, the naive estimated abundance is 360. However, one quadrat contained 24 animals, far more than the others. That patchiness should be discussed rather than hidden by a simple mean.
If the sample locations were chosen because the observer could already see many animals there, the 360 estimate may be biased upward. If they were randomly distributed, the estimate may be useful with substantial uncertainty. A translated report should say how quadrat locations were chosen. Multiplying a local density by the entire surface area is a model requiring a spatial assumption, not a universal truth.
Transfer problem: settlement on unexpected surfaces
A hatchery installs 12 biofilm plates and counts 700 new juveniles on them. A later inspection reveals another 150 attached to vessel walls and 50 on the outlet screen. If categories do not overlap, total visible settled individuals number 900. The plate-only share is 700/900×100, or 77.78%. It is correct to describe plates as the dominant observed settlement surface, but not to claim that 700 is the whole-tank settled count.
The outlet-screen animals may not survive an equipment transfer. Their presence is an observation, not proof of long-term retention. When compiling the next census, track whether they were safely moved, retained, detached or no longer observable. Without those movements, a later decrease in plate animals cannot be interpreted as mortality across the entire nursery.
Transfer problem: distinguish percentage points from percent growth
A conditioning treatment produces 20% settlement and an alternative produces 30%. The increase is 10 percentage points, or 50% relative to the original 20% rate. If a summary says ‘30% greater settlement’, it would correspond to 26% under the relative interpretation, not 30%. The language must specify either the new level or the calculated change.
Now consider 80% settlement followed by 10% survival of those settled animals. The resulting retained fraction of the originally introduced larvae is 8%, not 70% and not 90%. Improvement in an early response can be economically unimportant when later survival is very poor. This is why a settlement-cue study should not be portrayed as a finished solution to nursery production.
Transfer problem: time windows can change apparent success
A tank is checked twelve hours after exposure and 400 animals meet a settlement definition. Another observer checks after seventy-two hours and sees 650. The later count might reflect continued settlement, increased detectability, altered classification, or a combination. A higher number does not automatically mean improved biological success if the two observations differ in protocol, time and animal handling.
Reports comparing different treatments should use synchronized census schedules or justify adjustments. A treatment that induces rapid early attachment may rank highest at twelve hours but not at a later viable-juvenile checkpoint. Translate the observation window exactly. Do not place results from different timepoints side by side under a heading that implies a common endpoint.
Transfer problem: a screen transfer and an unclosed juvenile ledger
An initial census estimates 5,000 settled juveniles. During plate relocation, 400 visible individuals are documented as lost on handling surfaces, 3,900 are counted after transfer and 200 are counted in temporary holding containers. The recorded total after transfer is 4,100; including the 400 separately observed losses accounts for 4,500 of the initial 5,000. The remaining 500 are unaccounted within the evidence provided.
It is inappropriate to announce a 22% mortality rate simply because the receiving tank contains 3,900. Some individuals may remain in holding, others were documented lost, and the remainder lack confirmed fate. Depending on the intended reporting boundary, observed receiving-tank retention is 3,900/5,000=78%, while tracked retained animals across destination and holding represent 4,100/5,000=82%. Those are descriptive recovery fractions, not necessarily measured biological survival.
Transfer problem: a size threshold changes the accepted count
A nursery has 1,000 observed live juveniles at a transfer checkpoint. Of these, 720 meet the minimum measured size required for a particular experimental outplanting design; 280 remain smaller. The eligible fraction is 72% of observed live juveniles. If the original settlement cohort contained 4,000, the 720 eligible animals represent 18% of the original settled number, assuming the cohorts are linked and counting is complete.
Neither percentage tells us whether the animals will persist in the field. The size criterion is project-specific, and its measurement technique must be recorded. A sentence saying ‘72% survival to outplanting’ falsely treats a size-selection percentage as a survival outcome. Prefer ‘72% of live nursery animals at this checkpoint met the stated size criterion’.
Transfer problem: compare two commercial-scale methods
Method X yields an apparent 65% settlement from 10,000 competent larvae and 20% survival of the settled juveniles after thirty days. Method Y yields 45% settlement but 40% survival of those settled. Under these invented linked-cohort assumptions, X retains 10,000×0.65×0.20=1,300 juveniles, while Y retains 10,000×0.45×0.40=1,800. Thus the lower initial settlement method produces more living juveniles at the defined later checkpoint.
That example demonstrates why an unusually effective cue in a short assay may not represent an overall production improvement. Before recommending one method, a real hatchery would also need uncertainty estimates, replicate tanks, cost, growth, welfare, maintenance burden and longer follow-up. The English summary should report the initial and later endpoint side by side, with the final retained count calculated explicitly.
Transfer problem: a field release and incomplete detection
A conservation team releases 10,000 juvenile urchins and later sees 2,000 during a survey. It is incorrect to declare that exactly 8,000 died. Some survivors may be concealed, emigrated within the surveyed habitat, or simply not detected by the sampling procedure. Without a validated observation model, 20% is the observed count relative to releases under that search, not a direct estimate of survival.
If a properly designed study estimates detection probability and movement using suitable field methods, it may be possible to estimate abundance or survival with uncertainty. The estimate must be described as model-based, and assumptions about closure, habitat and detection should accompany it. A bilingual release report should separate the count released, count observed, area searched, detection method and inferential estimate.
Transfer problem: the initial cohort is measured twice
A settlement run records 4,000 larvae introduced, then another 2,000 added three days later. A final census finds 1,200 juveniles. If no animals were removed, an aggregate 20% final retained fraction relative to the six thousand introduced can be calculated. But this is not the survival percentage for either introduction cohort, because their exposure durations differ and the juveniles may not be attributable to an introduction date.
A report presenting the 1,200 as ‘30% settlement’ by dividing only by the first 4,000 excludes the later input and misstates the aggregate. However, using 6,000 does not recover the missing cohort histories. Identify both introductions and their dates, and describe the 20% as a combined apparent endpoint fraction rather than an age-matched survival estimate.
Frequently asked questions about juvenile and plate counts
Does settlement always happen on a plate? No. Animals may settle on walls, floors or other suitable surfaces depending on the system. Does a large settled count prove high survival? No. Later mortality can change the outcome greatly. Can biofilm condition be reduced to the word ‘clean’? No. The composition and intended role of biofilm require a defined method. Is a quadrat extrapolation a full count? No. It is an estimate under sampling assumptions.
Can the absence of visible juvenile animals prove they all died? Not when detectability is imperfect. Can field release numbers be added across several years? They can describe cumulative release events if records avoid double counting, but not necessarily the number currently alive. Why are experimental plates not always independent replicates? Because multiple plates in one vessel share the same water and treatment conditions. What matters is how units were assigned to treatments.
The independent verification workflow
Start by checking the definition of the endpoint in the methods. Mark whether a count refers to competent swimming larvae, initial attachment, morphologically completed metamorphosis, living juvenile urchins, transfer-eligible size, or field-observed animals. Then check timepoint and spatial boundary. A plate-only photograph and a full-container count are not equivalent evidence.
Next write every relevant denominator directly beside the percentage: introduced larvae, competent larvae, settled individuals, metamorphosed juveniles, or live animals at a previous census. Trace each intentional transfer, split or stock addition. Recalculate estimated totals from actual quadrat areas, surfaces and sampling assumptions. Check independent tanks and technical plate subsamples separately. This is more valuable than adding descriptive adjectives to a result.
End by stating what remains uncertain. Some animals may be hidden; a census may be incomplete; a trial may not have a comparable control or sufficient independent replication; an improvement at settlement may disappear later. Present those limits where the corresponding numerical claim appears. A technical translation is successful when a separate reader can reconstruct both the calculation and the biological meaning without inventing missing data.
An example of writing a cautious but useful conclusion
Suppose a newly completed nursery trial reports two candidate settlement treatments. Treatment A achieves 75% of introduced larvae meeting the settlement definition, and treatment B achieves 55%, but after a three-week follow-up both produce approximately the same number of visible live juveniles per initial thousand larvae. The evidence supports a strong short-term settlement response for A and no established advantage in the stated later output measure. The translation should not call A the winner of juvenile production merely because its first percentage is higher. Later survival may be lower, or its apparent count may be influenced by detection and sampling conditions.
A defensible result paragraph might say: ‘The treated surfaces increased initial observed settlement relative to the comparison condition under this trial’s scoring method. However, that advantage did not translate into a clearly greater number of observed live juveniles at the three-week checkpoint. The outcome is specific to the tested species, tank conditions, surface preparation and sampling design.’ This wording identifies both the positive early finding and the limiting late finding. It does not assert that the cue harms urchins, since the study may not have isolated the causal explanation for the later difference.
Finally, explain which information would allow an independent reader to assess practical value: number of independent tanks; larvae introduced and their competence; density and plate area; early and later census definitions; repeat counts and detection limits; survival intervals; maintenance inputs; and animal size at transfer. A commercial or conservation decision needs evidence at its actual final endpoint. When reports stop at settlement, the right next step is to ask for nursery follow-up, not to silently assume success. This is the crucial distinction between translating an experiment’s exciting first observation and interpreting its enduring biological result.
A final numerical cross-check
A final numerical cross-check can expose a hidden denominator mismatch. If 2,400 competent larvae were introduced, 1,200 animals settled, 900 were scored as metamorphosed and 450 remained alive at the nursery checkpoint, the settlement fraction is 50%, metamorphosis conditional on settled is 75%, and later survival conditional on metamorphosis is 50%. The cumulative live-juvenile fraction is 450/2,400, or 18.75%. The calculation 0.50 × 0.75 × 0.50 reproduces 0.1875. Report this only as a linked-cohort demonstration, because real census counts are often estimated and animals may be lost from view. A translator who replaces the 18.75% with the attractive headline ‘75% success’ has confused an intermediate conditional endpoint with the final output. A careful table will print the numerator, denominator and census stage for every percentage, so the choice cannot be made accidentally.
Primary sources and topic neighbours
James Cook University: sea urchin settlement cues, macroalgae and biofilms
Southern Cross University: settlement induction versus later survival