Quartz 5

An artificial productive ecosystem based on a fish/bacteria/plant association. 1. Design and management

Properties1
tagsmeta/type/exploratory, meta/region/europe, meta/fish/tilapia, meta/plant/tomato

34 min read

An artificial productive ecosystem based on a fish/bacteria/plant association. 1. Design and management

Metadata

  • Cite key: quillereArtificialProductiveEcosystem1993
  • Item type: Journal Article
  • Authors: I. Quillère, D. Marie, D. Roux, F. Gosse, J.F. Morot-Gaudry
  • Affiliation: Laboratoire du Métabolisme, INRA, Route de St-Cyr, 78026 Versailles Cedex, France (Quillère, Roux, Gosse, Morot-Gaudry); Installations Piscicoles Expérimentales (IPE), INRA, Domaine de Vilvert, 78352 Jouy-en-Josas Cedex, France (Marie)
  • Journal: Agriculture, Ecosystems & Environment 47 (1993) 13-30
  • Date: 10/1993
  • Date added: 2026-08-09
  • DOI: 10.1016/0167-8809(93)90133-a
  • Funding: Not stated. The Acknowledgements section (p.29) thanks named individuals for technical assistance, project initiation, encouragement, and translation, but names no funding body or grant.
  • URL: https://doi.org/10.1016/0167-8809(93)90133-a
  • PDF: Quilleré et al. - 1993 - An artificial productive ecosystem based on a fish.pdf

Opinion

A methods-and-design paper, not a results paper, and it reads that way: one un-replicated pilot system, one production cycle, no statistical test, and the authors say so themselves (“in this first trial no attempt was made to improve the animal production”). Its value is the engineering detail — three physically separated circuits (fish/biofilter loop, intermittent NFT hydroponic loop, capillary mineral-complement loop), explicit flow rates, and a genuinely useful worked example of how a fish-effluent-only nutrient solution falls short of a standard hydroponic recipe (Table 1) and how the authors patched the gap with a small liquid P/K supplement rather than the more common solid fertilizer or foliar-only approaches. The nitrogen story (stable, low water NO3 throughout most of the tomato cycle) is the paper’s real finding, but Table 1’s one dated water-mineral snapshot (Day 22, NO3 4.1 mmol/L) does not match the paper’s own later narrative claim that NO3 stayed in the 0.5-1.5 mmol/L range across essentially that same window — see Extraction notes, this is the paper’s one real internal conflict and it lands on a number central to the abstract’s headline claim. No replicate systems, no dispersion (SD) anywhere, and fish performance (growth, survival, FCR) is entirely deferred to the companion 1995 paper — treat this one as a system-design reference and a single descriptive data point, not as evidence for aquaponic-vs-hydroponic yield or fish-performance comparisons.

Abstract

An artificial ecosystem integrating three biological compartments (fish, bacteria, plants) in a closed system was developed with the aim of associating fish production with a vegetable crop purifying the fish water. The nitrogenous compounds excreted in dissolved form by the fish, and transformed by the bacteria, provide nitrogen nutrition for the plants. This association has the double advantage of savings in water for fish culture and the recycling of fish excretion as the main source of minerals for producing edible plants.

A pilot system of 2 m3 was set up for intensive animal and plant production and installed in a greenhouse to enable continuous production throughout all seasons. The fish chosen were tilapias (Oreochromis niloticus) and the plants were tomatoes (Lycopersicum esculentum) grown according to the nutrient film technique in recirculating hydroponics. A granular filter bearing the nitrifying bacteria was inserted between the fish tank and the plants. The system’s design was aimed at optimizing the functioning of the ecosystem, by the size of the different elements (fish tank, bacterial filter, hydroponic troughs) as well as by the choice of the recirculating water flow rate. During the first production cycle, we followed the evolution of the physico-chemical characteristics of the water and of the plant tissues, especially the nitrogen (NH3, NO2, NO3) and mineral compounds (K, Ca, Mg, SO4, PO4), in order to evaluate the functioning of the three compartments and to progressively develop the management of the plant compartment. The latter determined the overall equilibrium of the ecosystem by its capacity to absorb NO3 and NH4 in the recirculating water. The results were satisfactory as there was a stabilization of the nitrogenous compounds, in particular NO3, at a low level and a large plant production; in this first trial no attempt was made to improve the animal production.

This trial highlighted the main conditions to ensure the equilibrium of the ecosystem: size relationship between the three interacting compartments, dynamic management of the plant compartment (staggered crops) and the application of a mineral complement to obtain optimum plant growth. The nature of the mineral complement will depend on the composition of the water available on the production site.

Summary

The authors built and ran a single 2 m3 pilot aquaponic system at INRA Versailles (fish tank, granular nitrifying biofilter, and NFT hydroponic tomato troughs) over one ~105-day production cycle (1 September-15 December 1988), with three physically separate but interconnected water circuits so each compartment could be studied or disconnected independently. Sixty-nine tilapia (Oreochromis niloticus, progressively stocked to a total of 31.4 kg) supplied nitrogenous waste that was nitrified by a small granular biofilter and absorbed by 60 tomato plants (cv. Ferline) grown hydroponically in six (later eight) NFT troughs; a supplementary liquid mineral complement (initially a complete nutrient solution, later just acidified K/P) was fed directly to each plant pot via a third circuit because fish-effluent water alone was poor in potassium and phosphate and had an unfavourably high pH relative to a standard hydroponic recipe. Over the cycle, water NO3 was reported as staying low and comparatively stable, water pH declined gradually, and the tomato crop reached 82% of a comparison “conventional” crop’s yield with an almost identical fruit count but smaller average fruit size, attributed mainly to a lack of supplemental lighting late in the season and initial P/K/Fe deficiency symptoms that required foliar iron correction. Leaf and fruit mineral content at final harvest matched literature ranges for tomato reasonably well except for elevated phosphorus (attributed to intentional over-supply in the mineral complement) and mildly depressed potassium. No fish growth, survival, or feed-conversion data are reported here — the paper explicitly defers “the results of its functioning in varied conditions during a 2-year period” to a second paper (its own 1993b in-preparation reference), which was published as Quillère et al. 1995 in the same journal. The paper’s own contribution is the system’s engineering rationale and a first, largely qualitative account of how well the plant compartment kept the loop’s nitrogen balanced.

Experiment data

  • Location: INRA Laboratoire du Métabolisme / Installations Piscicoles Expérimentales, Versailles, France (greenhouse)
  • Design: Single pilot closed-loop system (no replicate systems, no hydroponic control system operated in parallel); descriptive monitoring of one production cycle. No randomization, no statistical test reported anywhere in the paper.
  • Replicates / n: 1 (one physical pilot system; hydroponic troughs increased from 6 to 8 during the season, but these are sub-units of the same single system, not independent replicate systems)
  • Duration: Day 0 (1 September 1988) to Day 105 (15 December 1988), the tomato crop’s full cycle; fish (69 adults, fully stocked by ~Day 45) continued in use beyond this paper’s own reporting window (confirmed by the 1995 companion paper — see Extraction notes)
  • Organisms: Tomato (Solanum lycopersicum) (cv. Ferline, called Lycopersicum esculentum in the paper) / Nile tilapia (Oreochromis niloticus)
  • Statistics: None. No ANOVA, no significance test, no dispersion (SD) reported for any measurement in the paper.
  • Nitrogen balance: Water NO3 reported by the Results text as “very low and relatively constant (between 0.5 and 1.5 mmol/L)” through most of the cycle, but Table 1’s single dated snapshot (Day 22) gives 4.1 mmol/L for the same water — see Extraction notes, ⚠️BLOCK.
  • Tomato yield: 0.75 kg fruit/plant (60 plants, 5 trusses each), 82% of a same-greenhouse “conventional crop” comparison, with an almost identical fruit count but 71% of its average fruit weight (p.25-26)
  • Leaf/fruit mineral content at harvest (Table 2, Day 105, % dry matter): Leaf blade N 3.4 / K 2.3 / Ca 7.1 / Mg 0.7 / P 1.1; Fruit N 2.7 / K 4.4 / Ca 0.3 / Mg 0.2 / P 0.7

Nitrogen balance and the NO3 conflict

This paper: The paper’s headline claim (Abstract, Conclusions) is that the plant compartment kept recirculating-water nitrate low and stable throughout the cycle, evidence that fish-excreted nitrogen alone can meet plant nitrogen demand without externally applied fertilizer nitrogen. The Results section states this explicitly for the main growth period: “From the introduction of the plants into the system … until their topping (Day 70), the functioning regime was stable: the NO3 content remained very low and relatively constant (between 0.5 and 1.5 mmol/L)” (p.22). However, Table 1 (p.17), captioned as the water’s mineral content “at Day 22,” reports NO3 = 4.1 mmol/L for the recirculating water — roughly 3-8x above the range the Results text describes as characteristic of that same window (tomatoes were transplanted into the troughs at Day 14, so Day 22 falls inside the “stable” period as described). See Extraction notes for the full ⚠️BLOCK writeup; this is the paper’s single most consequential unresolved figure, since it bears directly on the abstract’s own headline claim.

Compared with:

  • todo Lewis et al. 1981 — cited as prior work integrating plant crops into closed-circuit fish farming specifically to reduce water nitrate content (p.14)
  • todo Rakocy and Allison 1981 — cited for an aquaponic system using no fertilization at all (contrasted with this paper’s own liquid mineral-complement approach) (p.14, p.20)
  • todo Naegel 1977 — cited that 1.2 g/L NO3 (19 mmol/L) has no effect on fish growth, used to argue the system’s NO3 levels are far below any fish-toxicity concern (p.16)

System design and the three-circuit architecture

This paper: The system separated fish, bacteria, and plant compartments into three interconnected but independently operable circuits: Circuit I (continuous, 50 L/min, fish tank -> cyclonic decanter -> granular nitrifying biofilter -> cascade aerator -> back to fish tank), Circuit II (intermittent, 15-20 L/min depending on trough count, a derivation of Circuit I feeding the NFT hydroponic troughs and returning via a lift tank), and Circuit III (a liquid mineral complement delivered by capillary tube to each plant pot, active Day 25-89). The authors present this physical separation (rather than combining compartments in one tank, or using the plant growing medium itself as the biofilter, as some cited prior systems did) as a deliberate design choice enabling independent optimization and disconnection of compartments, and note their biofilter was downsized ~30% relative to a comparable published system (Watten and Busch, 1984) because the plants themselves absorbed part of the fish-derived NH4 directly.

Compared with:

  • todo MacMurty et al. 1990 — sand culture using recirculated aquacultural effluent, cited as an example of the alternative (single-tank, non-separated) design this paper argues against (p.20)
  • todo Zweig 1986 — integrated fish/hydroponic system where the plant growing substrate itself constitutes the biofilter, contrasted with this paper’s separate dedicated biofilter (p.20)
  • todo Watten and Busch 1984 — tilapia/tomato recirculating system used as the size benchmark for this paper’s ~30%-smaller biofilter (p.20-21)
  • todo Lewis et al. 1978 / Sutton and Lewis 1982 — cited as systems applying solid-form fertilization to hydroponic aquaculture crops, contrasted with this paper’s liquid mineral complement (p.20)

Tomato yield and mineral nutrition

This paper: The 60 tomato plants that completed the cycle to their fifth truss produced 0.75 kg fruit/plant on average, 82% of a “conventional crop grown at the same time in the greenhouse” with an almost identical fruit count but only 71% of that comparison crop’s average fruit weight; the authors attribute the shortfall mainly to reduced solar radiation for a late-season, unlit crop plus an initial 50% P/K leaf deficiency (foliar diagnosis, Day 48, relative to a hydroponic control) that was subsequently corrected by the Day-55-onward K/P mineral complement. Final-harvest leaf and fruit mineral content (Table 2) is described as broadly matching literature values for tomato except for elevated fruit/leaf phosphorus (140% of the literature reference, attributed to deliberate over-supply: the K/P complement provided ~4x the calculated phosphate requirement to ensure potassium adequacy) and a mild potassium shortfall (90% of reference in fruit, though the paper frames the ~25% depressed leaf K status as more limiting than the fruit K figure suggests, given potassium’s high within-plant mobility). Iron chlorosis appeared within ~10 days of transplanting and required foliar iron chelate correction plus weekly maintenance applications, attributed to the high water pH (7-8) rendering iron poorly soluble even after later inclusion in the mineral complement.

Compared with:

  • todo Statistical Analysis Systems 1969 — cited literature reference for tomato leaf/fruit mineral composition benchmarks (source of the 140%/90% comparison figures) (p.26) [secondary, cites SAS 1969]
  • todo Musard 1988 — cited alongside SAS 1969 as the literature basis for expected tomato mineral composition (p.26) [secondary, cites Musard 1988]
  • todo Odet and Musard 1989 — cited for the link between low potassium status and paler leaf colour in tomato, used to explain this system’s persistently pale leaves despite iron correction (p.25)

Linked claims

Citations to chase

  • todo Lewis WM, Yopp JH, Brandenburg AM, Schnoor KD (1981) — On the maintenance of water quality for closed fish production systems by means of hydroponically grown vegetable crops. In: Aquaculture in Heated Effluents and Recirculation Systems, Vol. I, Heenemann, Berlin, pp.121-129
  • todo Rakocy JE, Allison R (1981) — Evaluation of closed recirculating system for the culture of Tilapia and aquatic macrophytes. Proc. Bio-Engineering Symposium for Fish Culture, Vol.1, pp.296-307
  • todo Naegel LCA (1977) — Combined production of fish and plants in recirculating water. Aquaculture 10:17-24
  • todo MacMurty MR, Nelson PV, Sanders DC, Hodges L (1990) — Sand culture of vegetables using recirculated aquacultural effluents. Appl. Agric. Res. 5:280-284
  • todo Zweig RD (1986) — An integrated fish culture hydroponic vegetable production system. Aquaculture Mag. 12:34-40
  • todo Watten BJ, Busch RL (1984) — Tropical production of tilapia (Sarotherodon aurea) and tomatoes (Lycopersicon esculentum) in a small-scale recirculating water system. Aquaculture 41:271-283 (printed as “Warren, B.J.” in this paper’s own reference list — see Extraction notes)
  • todo Lewis WM, Yopp JH, Schramm HL Jr, Brandenburg AM (1978) — Use of hydroponics to maintain quality of recirculated water in a fish culture system. Trans. Am. Fish. Soc. 107:92-99
  • todo Sutton RJ, Lewis WM (1982) — Further observations on a fish production system that incorporates hydroponically grown plants. Prog. Fish Cult. 44:55-59
  • todo Quillère I, Roux L, Marie D, Roux Y, Gosse F, Morot-Gaudry JF (1995) — An artificial productive ecosystem based on a fish/bacteria/plant association. 2. Performance. Agriculture, Ecosystems & Environment 53:19-30 — the direct sequel to this paper, reporting fish growth/survival and multi-year performance data not covered here

Extraction notes

Relationship to Quillère et al. 1995 (companion paper, judgment call requested by task brief): This 1993 paper and pdfs/Quilleré et al. - 1995 - An artificial productive ecosystem based on a fish.pdf are an explicit two-part series on the same physical pilot system at INRA Versailles, not independent overlapping studies of different systems. This paper’s own Introduction states: “This paper deals with the design of the ecosystem and the management of the plant compartment… In a second paper (Quillère et al., 1993b), we will report the results of its functioning in varied conditions during a 2-year period” (p.14), and its Conclusions repeat “we studied the functioning of the system in other more or less favourable conditions and established nitrogen balances… (Quillère et al., 1993b)” (p.27). The reference list’s placeholder entry for “1993b” (“In preparation”) corresponds to the published 1995 paper, confirmed by opening the 1995 PDF directly: its Introduction states “In a preceding paper (Quillère et al., 1993) we have described an artificial ecosystem associating fish (tilapias), bacteria and plants (tomatoes)…” and its own title is explicitly “…2. Performance” against this paper’s “…1. Design and management.” This is not a case of an earlier preliminary report being superseded or duplicated by a later, more complete version of the same results — the two papers report different content (this one: system architecture, one single-cycle descriptive account, no fish performance data; 1995: fish growth/survival and a 2-year, multi-trial nitrogen-balance analysis across a spring-summer tomato trial and a separate winter lettuce trial). Both were extracted as separate trials.csv entries by design (this agent handled 1993 only, per instructions; the 1995 paper was assigned to a parallel agent). Flagging this explicitly as requested, since the two papers’ shared system and near-identical titles could otherwise look like accidental duplicate extraction.

Type classification judgment call: Recorded as exploratory. The paper reports data the authors collected themselves (water chemistry and plant tissue monitoring over a real production cycle), so it is primary research, not a review — but there is exactly one un-replicated pilot system, no randomized or paired treatment arms, and no statistical test anywhere (SCHEMA.md’s own test: “exploratory is descriptive and typically unreplicated with no formal hypothesis test… even if the authors call it a pilot” — which is literally the paper’s own self-description, “a pilot system of 2 m3 was set up”). This was a genuine judgment call against case-study (SCHEMA.md’s “single system or site described in depth”): the paper does report real repeated-measures quantitative data (Table 1, Table 2, Figs 3-6 time series) rather than a purely qualitative site description, which tipped the classification toward exploratory — the type that SCHEMA.md’s “which types produce CSV rows” section includes and case-study does not, and this paper’s Table 1/Table 2 numeric data is exactly the kind of trial-level data the CSV is meant to hold. Flagged here per CLAUDE.md’s instruction to note judgment calls rather than silently picking one.

⚠️BLOCK — Recirculating water NO3, Day 22 snapshot vs. stated “stable” range, p.17 (Table 1) vs. p.22 (Results). Table 1 caption (p.17): “Comparative pH and mineral content (mmol l-1) of the recirculating water (at Day 22)… Recirculating water [row]: pH 7-8, NO3 4.1, NH4 <0.1, K 0.15, Ca 2.75, Mg 0.6…” Results text (p.22), describing the period “from the introduction of the plants into the system… until their topping (Day 70)” — tomatoes were transplanted into the troughs at Day 14 (p.19: “introduced on 15 September (Day 14)”), so Day 22 falls inside this window: “the functioning regime was stable: the NO3 content remained very low and relatively constant (between 0.5 and 1.5 mmol l-1).” These two figures for the same parameter (recirculating-water NO3, mmol/L), overlapping in time, differ by a factor of ~3-8x, and nothing in the paper distinguishes them by compartment, method, or definition — both purport to describe the same recirculating water loop by the same presumed analytical method (nitration of salicylic acid, per Methods p.20-21). No textual signal favors one over the other: Table 1 is a specific dated measurement in a formal comparison table (normally the more reliable source), but it is a single-day snapshot, while the Results-text range is explicitly framed as characterizing the typical, stable condition of most of the growing cycle and is the number underpinning the paper’s central “low, stable NO3” claim (Abstract, Conclusions). Recorded UNCLEAR in the NO3-N cell; both candidate values preserved here and in the trial row’s Experimental Remarks. Affects: NO3-N cell only (no other cell derives from this passage); does not affect the qualitative “low/stable” narrative claim itself, only which specific number a reader might cite. Added to REVIEW.md.

WARN-MINOR — pH units typo, Biological requirements section, p.16. “pH between 5 and 11 (Balarin and Hatton, 1979), the most favorable for fish farming being generally between 6°C and 8°C (Billard and Marie, 1980)” — the second clause’s “°C” units are almost certainly a typesetting error for a continuation of the pH range being discussed (parallel structure with “pH between 5 and 11” immediately before), not an actual temperature statement; a genuine fish temperature range is given separately two sentences later (“optimal temperature between 24 and 35°C”). Not recorded in any cell (this is literature-cited background on tilapia biology in general, not this trial’s own measured pH or temperature, and no pHOptimal value was extracted from secondary citations at all — see below).

Metadata reconciliation, author name and initials. The PDF’s own byline (p.13, via pdftotext) reads “I. Quiller6 a’, D. Marie b, U Roux a, F. Gosse a, J.F. Morot-Gaudry a” — “Quiller6” is an OCR/encoding artifact for “Quillère,” and “U Roux” is most likely a scan/typesetting misread of a single-initial third author. Crossref (https://api.crossref.org/works/10.1016/0167-8809(93)90133-a, queried directly) gives the third author as D. Roux — distinct from the 1995 companion paper’s author list, which includes two Roux co-authors (L. Roux and Y. Roux, confirmed from that PDF’s own byline: “I. Quiller6 a’, L. Roux a, D. Marie b, Y. Roux a…”). Per CLAUDE.md (“PDF headers are frequently wrong… Crossref… is more reliable”), this paper’s author list uses Crossref’s “D. Roux,” which is plausible as the OCR source of the garbled “U Roux” (D and U are a plausible scan confusion) and is a third, distinct Roux from the two named in the 1995 paper. No existing vault note found for any of this paper’s five authors; all five are new wikilink targets.

[not reported] fields, grouped:

  • Fish: Fish size initial/final, FCR, SGR, Total Feed (kg), Fish biomass created (kg), Fish survival rate, Fish weight gain, Feed routine (feeds/day), Protein/P/K composition of feed, Initial Stock density (kg/m3) — this paper reports only the population-level stocking figures (69 adults, 31.4 kg total, progressively introduced over 45 days) and a daily-ration schedule (100->250 g commercial carp feed/day, 92% dry matter, 5.6% N); no individual fish weights, feed totals, survival/mortality, or growth outcome are given anywhere — fish performance is explicitly deferred to the companion 1995 paper (see above). NOT DERIVED even though a stocking density (31.4 kg / 2 m3) and mean individual weight (31.4 kg / 69 fish) are both arithmetically available from the given totals.
  • Water: EC, NO2-N, Daily Water exchange rate, Water classification, pHOptimal — Methods states pH and “electrical conductivity” were measured on every sample (p.20-21) but no EC value is given anywhere in text or Table 1 (“data not shown” in effect, though the paper never uses that phrase); NO2 is discussed only qualitatively (a “detrimental” threshold of >14 µmol/L is cited as context, p.22, not a measured trial value) and only appears numerically in Fig. 4, a chart (not read into any cell, per the never-read-a-figure rule); no % water exchange rate is stated, only that evapotranspiration/cleaning losses were “automatically replaced” from a tap-water-fed buffer tank (p.17). pHOptimal left NR rather than populated with the literature-cited tilapia pH tolerance range (5-11, “favorable” 6-8, Balarin & Hatton 1979 / Billard & Marie 1980, p.16) — these are secondary citations describing tilapia biology in general, not this system’s own stated operating target, and are preserved in the Compared With / Citations to chase sections instead.
  • Plant: Plant height, Leaf count, Plant dry matter (%), SPAD, Tissue nitrate AP, Days Plant after transplant, Plants/m2 — height, leaf count, a whole-crop dry-matter %, and SPAD/chlorophyll-meter readings are never reported (foliar diagnosis was via wet chemistry on ground leaf tissue, not a chlorophyll meter); tissue nitrate specifically was not assayed (only total Kjeldahl N, which is not nitrate); Days Plant after transplant is NOT derived from the two given absolute day-numbers (transplant Day 14, final harvest Day 105) per the no-derivation rule, even though both inputs are stated; Plants/m2 is not derivable because trough floor area is never given (only “ten plants, 50 cm apart” per trough and a trough count of 6-8).
  • Site: Lat/Long — the paper names only the city (Versailles, France) and never states coordinates in any form; per the prime directive, coordinates were NOT filled from outside geographic knowledge of where Versailles is.
  • Average room Temperature — only a winter heating setpoint (18°C) and a qualitative “ventilated in summer” statement are given (p.19-20); no measured trial-mean greenhouse air temperature is reported. Recorded in the cell with this setpoint caveat rather than left fully NR, since it is the paper’s own explicitly stated management figure, distinct from Water temperature (28°C, the fish-tank/root-zone water, separately and explicitly stated as “maintained”).

NO COLUMN items (full detail in the trial row’s Experimental Remarks): Table 1’s full water-mineral panel beyond nitrogen species (K, Ca, Mg, SO4, P, mmol/L, both the recirculating water and the standard-nutrient-solution reference recipe); Fig. 5’s narrative-described water K/Mg/Ca/SO4 trends over the cycle (K “increased slightly then disappeared almost completely,” Mg “decreased… stabilized around 0.35 mmol/L,” Ca “2.5-3 mmol/L,” SO4 “1.3-1.8 mmol/L” — all given as text ranges, not read off the chart itself); the foliar-diagnosis time series expressed as % of a hydroponic control (Fig. 6, Day 48 “50% P and K deficiency,” Day 75 “P requirements largely covered, K deficiency persisted” — relative percentages with no absolute units, not a plant.csv-compatible analyte value); the 6.5 kg total dry-matter figure for non-edible plant residue across all 60 plants (p.26); the “82% of conventional crop yield, ~71% average fruit weight, near-identical fruit count” relative comparison against an unspecified “conventional crop grown at the same time in the greenhouse” (p.25-26) — no absolute yield value is given for this comparison crop, so it could not be entered as a HYD-column value without derivation (see Trial structure below); the literature-cited tilapia biological tolerance ranges (pH 5-11, temperature 24-35°C lethal below 12-13°C, DO critical threshold 0.1 mg/L, ammonia/nitrite LC50/TLm figures, p.15-16) [all secondary, cited to Balarin & Hatton 1979, Huet 1970, Chervinsky 1982, Redner & Stickney 1979, Konikoff 1975, Naegel 1977]; Nitrobacter oxidation stoichiometry (>=13 mol NO2 per mol CO2 reduced) [secondary, cites Aleem 1970].

Trial structure and the “conventional crop” comparison. Recorded as a single trials.csv row (T1). This paper has no operated hydroponic control system: Table 1 compares the recirculating water’s composition only against a literature recipe (Lesaint-Coïc standard nutrient solution, 1983), not an actively run parallel hydroponic system, and the only mention of an actual simultaneously-grown comparison crop (“a conventional crop grown at the same time in the greenhouse,” p.25) gives no absolute yield figure, no stated growing system (soil vs. hydroponic vs. something else), and no replication or measurement detail — only three relative percentages (82% yield, ~71% average fruit weight, “almost identical” fruit count). Per SCHEMA.md’s NA convention (“Use… for the HYD columns when there is no hydroponic control”), the HYD-side columns are recorded NA throughout, and the relative comparison-crop percentages are preserved in Experimental Remarks under NO COLUMN rather than back-calculated into an implied absolute HYD value (which would be derivation from a percentage plus this trial’s own AP value).

Tags judgment call: Tagged Meta/Fish/Tilapia (Oreochromis niloticus, the paper’s sole aquaculture species) and Meta/Plant/Tomato (Lycopersicum esculentum / Solanum lycopersicum, the sole crop). Meta/Region/Europe per France. No new tag facets introduced; no TAGS.md file was found in the vault root to check against (see note below) — reused the existing Meta/Type/, Meta/Region/, Meta/Fish/, Meta/Plant/ facets already visible across other notes in notes/.

No TAGS.md found in the vault root at the time of this extraction (only CLAUDE.md, REVIEW.md, SCHEMA.md, TEMPLATE.md are present alongside notes/, pdfs/, out/). Tag facet reuse above was verified instead by grepping existing notes directly for Meta/Type/, Meta/Fish/, and Meta/Plant/ usage.

PDF quality: Real, usable text layer throughout (18 pages via pdftotext -layout), not a garbled OCR scan requiring NEEDS_OCR.md treatment. Body running text is fully legible aside from scattered, inconsequential character substitutions in accented names and running headers (e.g. “Quiller6,” “QuillerP,” “Quillerb” for “Quillère” across different page headers — an encoding artifact of the accent, not a content problem, since the byline and Crossref both confirm the correct form). Table 1 (p.17) and Table 2 (p.26) both parse cleanly into their stated columns once column headers are matched to the (fewer) numeric tokens per row — verified against the paper’s own running-text descriptions of the same figures (e.g. Table 2’s fruit P/K percentages cross-check against the “140%”/“90%” comparison sentence on p.26). Figures 1-6 are visual charts (schematic diagrams and time-series line/bar plots); their extracted text is heavily scrambled by the PDF’s original two-column-figure layout and was not used as a data source anywhere in this note or the CSVs, consistent with the never-read-a-figure rule — all figure-derived context here comes from the paper’s own running-text descriptions of what those figures show, not from parsing the figures’ extracted text.

Source: Quilleré et al. - 1993 - An artificial productive ecosystem based on a fish.pdf

Data Tables

Structured data extracted from this paper into the vault's trials.csv / plant_measurements.csv datasets. Fields the paper didn't report are omitted. Download the full datasets (measurements).

Trial Parameters

quillereArtificialProductiveEcosystem1993-T1

Fish

FieldValue
FishTilapia (Oreochromis niloticus)
Fish CategoryAdults (p.17-18)
N5.6
Feed regimeCommercial carp feed (92% dry matter, 5.6% N); daily ration increased gradually from 100 to 250 g/day over the cycle (p.18, p.22)
Fish trial duration (days)105

Water

FieldValue
Water recycle50 L/min (Circuit I, continuous)
Water volume in the system2000 (fish tank only; see remarks)
Water typeTap water, automatically replenished to replace evapotranspiration/cleaning losses (p.17)
Aq pH7-8
Dissolved Oxigen7.5 (82% saturation; plant-trough outlet, see remarks)
Water temperature28
TAN / NH4-N<1.4

Plant

FieldValue
PlantTomato (Lycopersicum esculentum), cv. Ferline
DetailsIntroduced Day 14 (15 Sept 1988) at first-flower-cluster stage; topped after fifth cluster on Day 70 (10 Nov); fruit picking began Day 97 (7 Dec); whole crop removed Day 105 (15 Dec 1988) (p.18-19, p.25)
Plant fresh weight750

System & Setup

FieldValue
System typeNutrient film technique (NFT), recirculating hydroponics (p.13); no substrate in the growing troughs
Media DetailsNone in main troughs (NFT thin film, 2-3 mm depth); pre-transplant nursery-stage seedlings grown in expanded clay-peat mix pots (p.19)
Biological system already in useY (Fixed aerobic granular-bed biofilter, 50 kg BIOGROG clay granulate (10-20mm), colonized by nitrifying bacteria (Nitrosomonas, Nitrobacter), self-loaded naturally (not inoculated) upon fish introduction (p.17-18, p.20))
Air supplementY (Cascade aerator at the end of Circuit I reoxygenates water returning to the fish tank; fish tank further oxygenated by a separate mechanical aerator (p.17-18))
Iron supplementedY (Foliar iron chelate spray (1.2 mg/L metal Fe) applied ~10 days after transplant to correct chlorosis; weekly reapplications; iron later also included in the liquid mineral complement (p.25, p.28))
RemineralizationY (Liquid mineral complement via Circuit III capillary tubes to each plant pot, Day 25-89 (26 Sept-29 Nov); Day 25-~54: complete Coic-Lesaint nutrient solution; Day 55+: limited to P and K as KH2PO4 acidified with H3PO4 to pH 4.5, 10 L/day, providing ~75% of theoretical K requirement (10 mmol/L) and ~4x the calculated P requirement (14.4 mmol/L) (p.24-25))
pH BuffersY (Mineral complement itself acidified to pH 4.5 (H3PO4) to prevent calcium phosphate precipitation before reaching plant roots; the main recirculating water loop was NOT actively pH-corrected and declined naturally over the cycle (Fig.3) from fish/plant CO2 respiration, biofilter activity, and the acidic complement (p.24-25))
Climate controlY (Greenhouse heated to 18 degC in winter, ventilated in summer; no further setpoints given (p.19-20))
Artificial LightingN (Explicitly stated no supplemental lighting used; late-season low solar radiation cited as a yield-limiting factor (‘a crop grown late in the season without extra lighting’, p.25-26))
Nutrient supplementedY (Same liquid mineral complement as Remineralization above, delivered directly to each plant’s root zone via Circuit III (p.20-21, p.24-25))
EquipmentClark-type oxygen electrode (Clark 1956); Nessler colorimetry or Parnas-Wagner microdistillation (ammonia); atomic absorption spectrophotometry (Ca, Mg); flame photometry (K); Kjeldahl method, Kjeltec System 1026 Tecator (total N); nitration of salicylic acid, Cataldo et al. 1975 (nitrate); diazotization of sulphanilamide, Rodier 1984 (nitrite); vanadomolybdic complex method (phosphorus); gravimetry of barium sulphate (sulphates); submersible circulation pumps; cascade + mechanical aerators; electric water heater; BIOGROG clay granular filter media (p.20-21)
Control ParametersFish tank water temperature maintained 28 degC via electric heater; greenhouse heated 18 degC winter/ventilated summer; mineral complement pH adjusted to 4.5; Circuit I flow 50 L/min continuous; Circuit II flow 15-20 L/min intermittent depending on trough count
CombinationTilapia and tomato; single-system fish/bacteria/plant closed loop, no hydroponic control system operated; NFT

Site

FieldValue
RegionEurope
CountryFrance
Average room Temperature18 (winter heating setpoint only, see remarks)

Results & Statistics

FieldValue
Measured Unitg/plant (yield); mmol/L (water minerals); % dry matter (plant tissue minerals)
Statistic DetailsNone reported (p.13-30)
Statistically analysedN
Replicates (n)1 (single pilot system; troughs increased 6->8 during the season, not independent replicate systems, p.19)
AP750

Experimental Remarks: TRIAL DEFINITION: T1 = the single pilot aquaponic system (2 m3 fish tank, granular nitrifying biofilter, NFT hydroponic tomato troughs), the only treatment arm this paper reports over one ~105-day production cycle (Day 0 = 1 Sept 1988 to Day 105 = 15 Dec 1988). There is no operated hydroponic control system: Table 1 (p.17) compares the recirculating water’s mineral profile only against a literature standard-nutrient-solution RECIPE (Lesaint-Coic type, 1983), not a parallel-run control system, and the Results section’s sole mention of ‘a conventional crop grown at the same time in the greenhouse’ (p.25) gives only relative yield percentages (82% of that crop’s yield, ~71% of its average fruit weight, ‘almost identical’ fruit count) with no absolute figures, growing-system description, or replication detail. HYD-labelled cells are therefore NA throughout this row per SCHEMA.md’s convention for ‘the HYD columns when there is no hydroponic control’; the relative comparison-crop percentages are preserved below under NO COLUMN rather than back-calculated into an implied absolute HYD value (would be derivation). | WARN-BLOCK Recirculating water NO3, Day 22 snapshot vs. stated ‘stable’ range: Table 1 (p.17), captioned ‘at Day 22’, gives Recirculating water NO3 = 4.1 mmol/L. Results text (p.22), describing the period from tomato transplant (Day 14, p.19) until topping (Day 70) — a window that contains Day 22 — states ‘the functioning regime was stable: the NO3 content remained very low and relatively constant (between 0.5 and 1.5 mmol l-1)’. These two figures for the same parameter, overlapping in time, differ by a factor of ~3-8x; nothing in the paper distinguishes them by compartment, method, or definition (same recirculating water loop, same nitration-of-salicylic-acid method, Methods p.20-21). No textual basis favors one over the other: Table 1 is a specific dated table value (normally preferred) but is only a single-day snapshot, while the Results range is explicitly framed as the trial’s typical/stable condition and underlies the Abstract’s headline ‘stabilization… at a low level’ claim. Recorded UNCLEAR in the NO3-N cell; both candidates given here for the record. Affects: NO3-N cell only. Added to REVIEW.md. | UNIT CONVERSION ONLY: NH4 <0.1 mmol/L (Table 1, Day 22 recirculating water) -> <1.4 mg/L NH4-N (x14 N molar mass). NO3 candidates converted the same way: Table1’s 4.1 mmol/L -> 57.4 mg/L NO3-N; Results text’s 0.5-1.5 mmol/L range -> 7-21 mg/L NO3-N (both preserved in the UNCLEAR cell’s basis, see WARN-BLOCK above). Plant fresh weight 0.75 kg/plant (p.25) -> 750 g/plant. Coordinates: none given anywhere in the paper (city name ‘Versailles’ only) — Lat/Long left NR rather than filled from outside knowledge of the city’s location. | NOT DERIVED, left NR: Initial Stock density (paper gives 69 fish, total 31.4 kg, 2 m3 tank — kg/m3 not stated); Fish size initial/final (only the 31.4 kg population total given, no mean per-fish weight); FCR, SGR, Total Feed (kg), Fish biomass created (kg), Fish survival rate, Fish weight gain (fish performance is explicitly deferred by this paper to its companion 1995 paper — see note file); Days Plant after transplant (transplant Day 14 and harvest Day 105 both stated, but their difference (91) is not itself stated and was not computed); Plants/m2 (10 plants/trough, 50 cm apart, 6-8 troughs given, but trough floor area never stated, so no density is derivable without an assumed trough width). | NO COLUMN: Table 1 full water-mineral panel beyond N species, mmol/L, both recirculating water and the standard-nutrient-solution reference recipe — K 0.15/5.2, Ca 2.75/3.1, Mg 0.6/0.75, plus the cation/anion K-Ca-Mg exchange columns and H2PO4 0.25/6.9, SO4 1.25/0.75 (recirc./standard respectively; p.17). Fig.5 narrative water-mineral trends (not read off the chart itself, only the paper’s own text description, p.23-24): K ‘increased slightly [after fish introduction] then disappeared almost completely ~10 days after plants introduced’; Mg ‘decreased after 30 days of cultivation (Day 45), stabilized around 0.35 mmol/L’; Ca ‘2.5-3 mmol/L, varied slightly’; SO4 ‘1.3-1.8 mmol/L, varied slightly’. Foliar diagnosis time series as % of a hydroponic control (Fig.6, p.25): Day 48 ‘P and K deficiency equal to 50% of control’; Day 75 ‘P requirements largely covered, K deficiency persisted’ — relative percentages, no absolute units, not entered in plant.csv. Non-edible plant residue: 6.5 kg total dry matter across 60 plants, not recycled in this system (p.26). Comparison-crop relative yield (p.25-26): ‘a conventional crop grown at the same time in the greenhouse’ — AP represented 82% of its yield, ~71% of its average fruit weight, ‘an almost identical number of fruit’; no absolute value given for the comparison crop itself. Literature-cited tilapia tolerance ranges [all secondary, not this paper’s own data]: pH 5-11 (Balarin & Hatton 1979), favorable pH ~6-8 (Billard & Marie 1980, printed with a probable degC-for-pH unit typo, p.16 — see note file WARN-MINOR), lethal temperature below 12-13degC / optimal 24-35degC (Huet 1970), DO critical threshold 0.1 mg/L (Chervinsky 1982), NH3 48-h LC50 2.9 mg/L for Tilapia aurea (Redner & Stickney 1979), NO2 96-h TLm 24.8 mg/L for channel catfish (Konikoff 1975), NO3 1.2 g/L (19 mmol/L) with no effect on fish growth (Naegel 1977). Nitrobacter stoichiometry >=13 mol NO2 oxidized per mol CO2 reduced [secondary, cites Aleem 1970]. | Fish Category = ‘Adults’ (p.17/p.18: ‘a final charge of 69 adults’); a general species benchmark (‘marketable after 6-7 months with an average weight of 300-400 g’, p.15) is background biology, not this batch’s own measured weight, so not used as Fish size initial/final. | Feed regime: commercial carp feed, 92% dry matter, 5.6% N, daily ration increased gradually from 100 g/day to 250 g/day over the cycle as fish were progressively stocked (p.18, p.22); feeding stopped after Day 97. Feed routine (feeds/day) not stated. | Water recycle: Circuit I (fish tank <-> biofilter loop) = 50 L/min continuous (p.19); Circuit II (hydroponic loop) = 2.5 L/min per trough intermittently, i.e. 15-20 L/min total depending on the 6-8 troughs in use (p.19-20) — Circuit I’s figure used in the Water recycle cell as the main system recirculation rate, Circuit II’s given here for reference. Circuit III (mineral complement) is a 10 L/day capillary drip, not a recirculating flow. | Water volume in the system: 2000 L is the FISH TANK ONLY (p.17: ‘a fish tank (2 m3)’); a 150 L decanter is also sized, but the buffer tank (P1), biofilter, hydroponic troughs, lift tank (P2), and mineral-complement reservoir volumes are not stated, so no system total is given. | Aq pH = 7-8: stated twice independently (Table 1, Day 22 snapshot, p.17; and again describing the troughs’ pH when the Day-55-onward K/P complement was formulated, p.24) — not a trial mean/SD. Fig.3 (chart only, not read into this cell) shows pH declining roughly continuously from ~8.3 at Day 0 to ~6.3 by Day 120; ‘7-8’ is the paper’s own descriptive mid-trial characterization, not a claim that pH was constant across the whole 105-day cycle. | Dissolved Oxygen = 7.5 mg/L / 82% saturation: paper states water O2 ‘varied between 56% of the saturation rate (5.2 mg/L when leaving the fish tank) and 82% (7.5 mg/L when leaving the plant troughs)’ (p.20-21); per SCHEMA.md’s per-compartment rule, the plant-bed/hydroponic-unit value (7.5 mg/L, plant-trough outlet) is used; the fish-tank-outlet value (5.2 mg/L, 56%) is the lower end of the same range, given here for reference. Neither is a trial mean+SD, both are stated range endpoints tied to a specific circuit location. | Water temperature = 28 degC: ‘the water temperature was maintained at 28 degC using an electric heater’ (p.18) — an actively maintained/achieved setpoint, not merely nominal, and matches the root-zone temperature statement (p.17, ‘determined by the temperature of the fish tank (28 degC)’). No SD given (no stated variability around this maintained value). | TAN/NH4-N: Table 1, Day 22, recirculating water NH4 <0.1 mmol/L -> <1.4 mg/L NH4-N (unit conversion, see above); single snapshot, no trial mean/SD given. | Average room Temperature = 18 degC (winter heating setpoint only, p.19-20: greenhouse ‘heated in winter (18 degC) and ventilated in summer’); not a measured trial-mean air temperature, and no summer setpoint is given (‘ventilated’ only). Distinct from Water temperature (28 degC, fish-tank/root-zone water). | Statistically analysed = N: no ANOVA, significance test, or dispersion (SD) is reported anywhere in the paper; a single un-replicated pilot system was monitored descriptively over one production cycle. | Type classification judgment call (exploratory vs. case-study), the paper’s explicit two-part relationship to its own companion 1995 paper (same system, different content — design/one-cycle description here vs. multi-year fish-performance data there), and full metadata-reconciliation notes (author byline OCR vs. Crossref) are in the note file’s Extraction notes section.

Plant Measurements

TrialSystemCategoryAnalyteValueUnitSig.Location
quillereArtificialProductiveEcosystem1993-T1APmineralN3.4% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralK2.3% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralCa7.1% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralMg0.7% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralP1.1% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralN2.7% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralK4.4% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralCa0.3% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralMg0.2% dry matterNRTable 2, p.26
quillereArtificialProductiveEcosystem1993-T1APmineralP0.7% dry matterNRTable 2, p.26

Graph View

Backlinks