The Evidentiary Record — Aquaponics Research, 1977 to Present

## 1. What the NCSU Research Established (1984–1997)

Between 1984 and 1990, a multi-disciplinary research group at North Carolina State University led by Mark McMurtry and supervised by Professor Douglas Sanders conducted three series of controlled experiments on recirculating integrated fish-vegetable culture [1–6]. The research consortium included 45 scientists, consultants, and technical resources spanning horticultural science, aquatic ecosystems, soil science, agricultural engineering, and economics. Ten held the title of Fellow in their respective scientific organizations (FAAAS, FASHS, FASABE). Two had direct ties to NASA’s Controlled Ecological Life Support System (CELSS) program: Dr. John C. Sager (FASABE) and Dr. Merle H. Jensen [7,8].

The third series — the ratio study — employed the following design:

**16 independent systems.** Each system was a fully closed recirculating loop: its own fish tank, its own sand biofilter, its own water chemistry, its own fish population. The 16 systems operated concurrently in the same greenhouse at Raleigh, North Carolina.

**4 treatments.** Four biofilter-to-tank volume ratios were tested: 0.67:1, 1.00:1, 1.50:1, and 2.25:1.

**4 replicates per treatment.** Four independent systems assigned to each ratio.

**3 temporal experiments.** The same 16 systems were used for a summer tomato trial (cv. ‘Laura’, 103 days, 1988), a fall fallow/cucumber trial (cv. ‘Fidello’, including a 42-day no-plant stress test, 1988), and a spring tomato trial (cv. ‘Kewalo’, 132 days, 1989) [1–6].

**Formal statistical analysis.** Randomized complete block design. ANOVA with Scheffé F-test. LSD values reported at P ≤ 0.05, 0.01, and 0.005 for all treatment means [1–4].

**Complete data reporting:**

– Fish: 362-day composite growth data, feed conversion ratios (1.29–1.51), standing biomass ceiling (12 kg/m³), monthly biomass equalization protocol [4,5]

– Feed: 14-element laboratory analysis of Purina Fish Chow 5140, manufactured without vitamin/mineral premixes to simulate developing-world conditions. N 4.65%, P 0.88%, K 1.20%, Ca 1.31%, Mg 0.28%, S 1,600 ppm, Fe 201 ppm, Mn 52 ppm, Zn 65 ppm, Cu 12 ppm, B 22 ppm, Mo 0.4 ppm [1,4]

– Plants: 14-element tissue analysis (leaf, fruit, whole plant) for two tomato cultivars across four biofilter ratios. Leaf N 4.30–4.62%, P 0.72–0.78%, K 3.65–3.88%, Fe 220–235 µg/g, Mn 143–211 µg/g, Zn 99–152 µg/g [1,2]

– Water: complete end-of-year irrigation chemistry, daily dissolved oxygen, ammonia, nitrite, nitrate, pH [4–6]

– Mass balance: full-system nitrogen mass balance (47–69% N recovery in fish + plant tissue). Nutrient assimilation as percentage of feed input, demonstrating mature-bed nutrient banking (K 131–201%, Ca 23–107%, Mg 83–323% of feed input) [1,2]

– Economics: caloric and protein production per unit water, projected revenue per composite unit area [3]

**Publications.** Results appeared in six peer-reviewed publications:

| Year | Paper | Journal |

|——|——-|———|

| 1987 | Mineral content and yield of bush bean, cucumber, and tomato cultivated in sand and irrigated with recirculating aquaculture water [Abstract] | HortScience 22(3): 510 |

| 1990 | Sand culture of vegetables using recirculated aquacultural effluents | J. Applied Agricultural Research 5(4): 280–284 |

| 1993 | Yield of tomato irrigated with recirculated aquacultural water | J. Production Agriculture 6(3): 428–432 |

| 1993 | Mineral nutrient concentration and uptake of tomato irrigated with recirculating aquaculture water | J. Plant Nutrition 16(3): 407–419 |

| 1997 | Effects of biofilter/culture tank volume ratios on productivity of a recirculating fish/vegetable co-culture system | J. Applied Aquaculture 7(4): 33–51 |

| 1997 | Efficiency of water use of an integrated fish/vegetable co-culture system | J. World Aquaculture Society 28(4): 420–428 |

The 1997 JWAS paper accumulated approximately 115 citations (publisher data) and was listed among the top 100 most-requested JWAS articles in January 2024. The total corpus has approximately 264–352 citations spanning 30–39 years, with recent citations in PMC/NIH (2025), MDPI Water (2025), Springer Nature Link (2024), and Nature Scientific Reports (2022) [9].

**Independent commercial replication.** In 1992–1994, a USDA-funded trial was conducted by Dr. Boone Mora, a retired veterinarian with no prior aquaponics experience, at the NCSU Horticultural Research Station. The system produced approximately 22,780 kg of fish per year (115.6 kg/m³/yr) and generated $30,000–50,000 in annual profit after expenses. Revenue per square meter was 7.5 times the UVI benchmark. The system was sold as an ongoing business [10,11].

—

## 2. What Came Before (1977–1986)

Three earlier publications described integrated fish-vegetable systems using recirculating water:

|——-|——|———|——–|——–|

Each was published in a peer-reviewed journal. Each demonstrated feasibility — that fish and plants can be grown together in recirculating water. None used replication, treatment comparisons, or inferential statistics. They demonstrated that something worked. They did not test why, how, or under what conditions [15].

The UVI system developed by James Rakocy beginning in 1987 was a single demonstration unit that evolved through multiple design changes over years of operation. Later reviewers described it as “demonstrative, not experimental” [15,16]. Its widely cited production data has never been replicated from independent, concurrent experimental units [16,17].

—

## 3. The Gap — 22 Years Without Properly Replicated Research (1997–2019)

### 3.1 The “House of Cards” Effect

A systematic review published in 2020 by Wilson Lennard analyzed 61 aquaponic studies from 2000 to 2020. The finding [17]:

> “Sixty-one percent (61%) of all the studies were deemed to have applied no or incorrect replication (no or incorrect replication: 56% of fully recirculating system studies, 100% of decoupled system studies, 86% of irrigated RAS water studies).”

This is not a marginal deficiency. In biological science, if you cannot replicate your experimental unit, you cannot isolate the treatment effect from random variation. When researchers treat multiple samples drawn from a single shared system as though they were independent replicates — measuring dozens of plants in one tank and running ANOVA as if each plant were an independent data point — they commit pseudoreplication, a fundamental statistical error described by Hurlbert (1984) [18]. The result is not merely “weaker” evidence. It is scientifically invalid inference [17].

The problem compounds over time. Early, highly cited papers — including the UVI demonstration unit and methodologically flawed comparative studies — became embedded as foundational benchmarks. Subsequent researchers built experimental designs on these unverified ratios. When those researchers in turn applied pseudoreplication to their own findings, published papers entered the consensus. Future work was built on compounded methodological errors. A 2022 assessment described this as a “house of cards” effect — a body of literature that appears robust from citation metrics but is structurally hollow [15,19].

### 3.2 The Only Pre-2015 Candidate

Lennard & Leonard (2004) [20] compared constant flow versus reciprocating flow in gravel-bed aquaponic systems. The study used 12 physically independent units, each with its own fish tank, biofilter, gravel bed, and pump. This met the criterion of system-level replication.

The study had a fatal design flaw, however: the constant-flow treatment produced continuous waterfall aeration while the reciprocating-flow treatment did not. Aeration — not flow regime — could explain any observed difference in performance. The two treatments differed on more than one variable, making it impossible to attribute effects to flow regime alone. Multiple independent analyses have identified this as a confounded experiment that produced misleading conclusions embedded in aquaponic practice [15,19].

### 3.3 Post-2015 Studies

After McMurtry stated his challenge (see Section 4), several studies achieved system-level replication. The first to do so without fatal design confounds was Maucieri et al. (2019) [21], which used 9 independent aquaponic units.

Prior to this, Silva et al. (2017) [22] claimed an RCBD with four blocks in a semi-intensive aquaponics system. However, only two complete aquaponics installations were built. The four plant tanks per treatment shared a common fish system, common water, and common recirculation loop — making them sampling units, not experimental units. The authors acknowledged this implicitly: fish tank parameters had “no statistical tests carried out due to lack of replicates” [22].

### 3.4 Independent Corroboration of the Gap

The methodological vacuum in aquaponics has been identified by multiple independent sources, not only by McMurtry:

– **Lennard (2020):** 61% pseudoreplication rate across the field [17].

– **Milliken (2022):** “When looking at the literature, we normally see very few or no replicates, or two replicates per treatment at the most” [15].

– **Yep & Zheng (2019):** Identified lack of replication and inadequate reporting as persistent problems in aquaponic research [23].

– **Pryce (2025):** Documented that aquaponic claims of viability are “often premature due to persistent gaps in the scientific approach,” including “insufficient rigorous evidence concerning economic performance” and “practical yield comparisons” [24].

– **Flett (2017):** Found that much aquaponics information “arises outside traditional academic channels,” “lacks peer review,” and may reflect “deliberate or accidental misinformation,” particularly when “authors might benefit financially from the systems or equipment they discuss” [25].

– **Nichols (2015):** Warned against “widespread investment in aquaponics” before rigorous comparative studies are available, cautioning about the “risk of adopting a potentially less productive or economically unviable technology” [26].

– **Greenfeld et al. (2019):** Found that “the actual profitability and net societal contribution of the aquaponics sector are still largely unclear” [27].

– **Goodman (2011):** Noted that literature “frequently details system construction and management from a scientific perspective but often neglects crucial analyses of financial viability” [28].

– **Goddek et al. (2015):** Identified “challenges of sustainable and commercial aquaponics” including the gap between claimed and demonstrated economic performance [29].

– **Palm et al. (2019):** In the Springer aquaponics textbook, identified DWC system limitations including “substantial labor, maintenance, and energy demands” that raise “questions about scalability and economic feasibility” [30].

### 3.5 The Gap in Years

– First published replicated research: McMurtry, 1990–1997 [1–6]

– Next study with proper whole-system replication and no fatal design confound: Maucieri et al., 2019 [21]

– **Gap: 22 years**

—

## 4. The Evidentiary Challenge (McMurtry, ca. 2014–2015)

In an online forum approximately 25 years after his experiments concluded, McMurtry made two statements that form the basis of the claim under examination.

**Statement 1 (MC-101-S19):**

> “To my/our knowledge, other than UVI’s dubious if not also deceptive self-reporting, no one has yet documented or disclosed any tangible findings resultant to any so-called AP ‘system’ — at all — NTM has anyone conducted actual scientific research in the past 25 years. I/we have made this point (challenge) repeatedly and to date no one has yet to question its veracity much less attempted to provide any evidence whatsoever that remotely contradicts said assertion.”

**Statement 2 (MC-100-S03):**

> “These were also randomized complete block design experiments with 4 spatial replicates each. TMK, these the first and the last studies in integrated aquaculture that employed replication (spatial and temporal) and statistical analysis.”

These statements contain three claims:

1. That the NCSU studies were the **first** in integrated aquaculture to employ replication and statistical analysis.

2. That they were the **last** to do so — i.e., no one had done it in the 25 years since.

3. That no one in that period had disclosed tangible findings from properly replicated experimental research.

**Verification of Claim 1 (first):** Supported. Every pre-McMurtry recirculating integrated aquaculture publication was a single-system demonstration without treatment comparisons, replication, or inferential statistics. The NCSU program was the first to apply RCBD, ANOVA, and LSD to this category of system [12–14,16].

**Verification of Claim 2 (last, at time of utterance):** Supported. At the time McMurtry wrote (ca. 2014–2015), the only study between his 1997 publications and his forum post that claimed system-level replication was Lennard & Leonard (2004) — a study with a fatal design confound that McMurtry explicitly rejected as legitimate scientific research [20]. Silva et al. (2017), Maucieri et al. (2019), and all subsequent properly replicated studies appeared years after his statement [21,22]. The first valid candidate was published 4–5 years after he spoke.

**Verification of Claim 3 (no tangible findings from properly replicated research):** Supported by the independent sources cited in Section 3.4. The Lennard (2020) 61% pseudoreplication finding, the Milliken (2022) assessment of “very few or no replicates,” and the multiple reviews identifying systemic methodological failures all corroborate that the published record at the time contained no properly replicated experimental research outside the NCSU program [15,17,19,23–30].

—

## 5. The Deeper Problem — What Research Cannot Answer Without Replication

The methodological failures in aquaponics are not an academic technicality. They have concrete consequences.

**Feed profiling.** Fewer than 0.5% of aquaponics papers conduct complete elemental analysis of fish feed [17]. Without knowing the exact elemental composition of the sole nutrient input, it is impossible to determine whether observed plant deficiencies reflect system failure, feed inadequacy, or nutrient antagonism in the water column. Most researchers rely on the manufacturer’s proximate analysis label, which provides only crude protein, fat, fiber, moisture, and ash — not the milligrams-per-kilogram values of iron, zinc, copper, boron, or molybdenum required for plant nutrition [31].

**N-P myopia.** Most “nutrient mass balance” studies restrict analysis to nitrogen and phosphorus. Plants require at least 14 essential elements. Tracking only N and P while ignoring potassium, calcium, magnesium, sulfur, iron, manganese, zinc, copper, boron, and molybdenum means that researchers cannot diagnose the cause of crop failure when it occurs — and cannot claim to have validated nutritional completeness when it doesn’t [31].

**Single-system n=1 comparisons.** As documented in the Graber & Junge (2009) case study, researchers have published comparative aquaponic vs. hydroponic performance claims based on a single aquaponic system (n=1) and a single hydroponic system (n=1), then applied ANOVA as if the individual plants within those systems constituted independent replicates. The statistical inference drawn from such studies — that one system outperforms the other — is invalid by the standards of biological experimental design [18,31].

**Data laundering.** In the same Graber & Junge (2009) publication, the authors incorporated tables and figures from external, unrelated, and in some cases non-peer-reviewed sources (student theses, separate publications) and presented them within the context of their own study in a manner implying they represented the study’s direct empirical results. When these external adaptations are stripped away, the only original experimental data consisted of raw NPK removal rates from a single unit — with no standard deviations, no error bars, and no p-values [31].

**The net growing area fallacy.** Commercial aquaponics proponents cite revenue projections based exclusively on planted raft area, ignoring the mandatory footprint of clarifiers, filters, sumps, degassing tanks, and perimeter access aisles. A facility claiming 214 m² of growing area may actually occupy 500–900 m² of total greenhouse floor space. Revenue-per-square-meter figures derived from the net area alone are mathematically inflated [32].

These problems are not fringe criticisms. They are documented in the peer-reviewed literature by the field’s own researchers.

—

## 6. Scope Comparison — What Has Not Been Repeated

| Parameter | McMurtry (NCSU, 1988–1997) | Has It Been Repeated? |

|—|—|—|

| Independent systems | 16 | No. Closest: 12 (Lennard 2004, fatal confound) or 9 (Maucieri 2019) |

| Treatments | 4 BFV ratios | No. Most studies test 2–3 treatments |

| Spatial replicates per treatment | 4 | No. Most replicated studies use 2–3 |

| Temporal experiments | 3 different seasons | No. Yang & Kim (2023) came closest with 3 temporal repeats |

| 14-element feed profiling | Yes | No. Lennard (2020): <0.5% of papers |

| 14-element tissue analysis | Leaf, fruit, whole plant | Rare |

| Full-system N mass balance | 47–69% recovery | Rare |

| Mature-bed nutrient banking data | K 131–201%, Ca 23–107%, Mg 83–323% of feed | No |

| LSD reporting for all means | Yes | Inconsistent |

| Interdisciplinary scope | 7 disciplines, 5 departments, 9 consultants, 20+ institutions | No |

| Independent USDA-funded replication | Boone Mora (1992–1994) | No |

—

## 7. Summary

1. **1977–1986:** Three peer-reviewed studies demonstrated feasibility of recirculating integrated fish-vegetable culture [12–14]. All were single-system demonstrations. None used replication, treatment comparisons, or inferential statistics.

2. **1988–1997:** The NCSU iAVs program was the first to apply proper experimental design: 16 independent systems, 4 treatments, 4 replicates each, RCBD, ANOVA, LSD, and 3 temporal experiments. Published across 6 peer-reviewed papers spanning aquaculture and horticulture journals [1–6]. Funded by USDA Special Grant No. 85-CRSR-2-2664. Independently replicated at commercial scale in a USDA-funded trial [10,11].

3. **1997–2019:** A 22-year gap. No study achieved proper whole-system replication without fatal design confounds. The only pre-2015 candidate [20] was confounded. The next candidate [22] was pseudoreplicated. The first valid study appeared in 2019 [21].

4. **2000–2020:** A systematic review by Wilson Lennard found 61% of aquaponics studies used no or incorrect replication [17]. Multiple independent academic sources have corroborated the field’s methodological deficiencies [15,19,23–30].

5. **Present day:** The NCSU program’s scope — 16 concurrent independent systems, 4 treatments with 4 replicates, 3 temporal experiments, complete 14-element feed and tissue profiling, full-system mass balance, transparent statistics, and independent commercial replication — remains unmatched.

When McMurtry stated in approximately 2014–2015 that his work was “the first and the last” properly replicated research in integrated aquaculture, the published record contained no counterexample. The 22-year gap between his last publication and the next valid study retroactively confirms the substance of the claim. The methodological vacuum he identified has since been verified by independent academic sources — including by the researcher whose own 2004 confounded study was the only pre-2015 candidate.

—

## References

1. McMurtry, M.R., D.C. Sanders, P.V. Nelson, and S.G. Nash. 1993. Mineral nutrient concentration and uptake of tomato irrigated with recirculating aquaculture water as influenced by quantity of fish waste products supplied. *Journal of Plant Nutrition*, 16(3): 407–419.

2. McMurtry, M.R., D.C. Sanders, R.P. Patterson, and S.G. Nash. 1993. Yield of tomato irrigated with recirculated aquacultural water. *Journal of Production Agriculture*, 6(3): 428–432.

3. McMurtry, M.R., D.C. Sanders, J.D. Cure, R.G. Hodson, B.C. Haning, and P.C. St. Amand. 1993. Food value, water use efficiency and economic productivity of an integrated aquaculture-olericulture system as influenced by component ratio. *HortTechnology* (ASHS), 3(4): 463. [Abstract]

4. McMurtry, M.R., D.C. Sanders, J.D. Cure, and R.G. Hodson. 1997. Effects of biofilter/culture tank volume ratios on productivity of a recirculating fish/vegetable co-culture system. *Journal of Applied Aquaculture*, 7(4): 33–51. DOI: 10.1300/J028v07n04_03

5. McMurtry, M.R., D.C. Sanders, J.D. Cure, R.G. Hodson, B.C. Haning, and P.C. St. Amand. 1997. Efficiency of water use of an integrated fish/vegetable co-culture system. *Journal of the World Aquaculture Society*, 28(4): 420–428. DOI: 10.1111/j.1749-7345.1997.tb00290.x

6. McMurtry, M.R., P.V. Nelson, D.C. Sanders, and L. Hodges. 1990. Sand culture of vegetables using recirculated aquacultural effluents. *Journal of Applied Agricultural Research*, 5(4): 280–284.

7. NCSU iAVs Research Group documentation. McMurtry personal archive. Interdisciplinary consortium included personnel from Horticultural Science, Botany/Crop Science, Zoology/Aquaculture, Biological and Agricultural Engineering, Soil Science, Statistics, and Agricultural Economics at NCSU, plus external consultants including Dr. John C. Sager (NASA CELSS) and Dr. Merle H. Jensen (University of Arizona, NASA consultant).

8. McMurtry, M.R. 1990. Performance of an integrated aquaculture-olericulture system as influenced by component ratio. PhD Dissertation. North Carolina State University, Raleigh, NC.

9. Citation data: ResearchGate profile and JWAS publisher counter (May 2026). See `W1-master-reference.md` for full citation analysis.

10. McClintic, D. 1990. Double Duty Greenhouse. *The Furrow* (John Deere), Jan–Feb: 16–17.

11. McClintic, D. 1994. Double-duty greenhouse. *The Furrow* (John Deere), Mar–Apr: 41–42.

12. Naegel, L.C.A. 1977. Combined production of fish and plants in recirculating water. *Aquaculture*, 10: 17–24. DOI: 10.1016/0044-8486(77)90029-1

13. Lewis, W.M., J.H. Yopp, H.L. Schramm Jr., and A.M. Brandenburg. 1978. Use of hydroponics to maintain quality of recirculated water in a fish culture system. *Transactions of the American Fisheries Society*, 107: 92–99.

14. Watten, B.J. and R.L. Busch. 1984. Tropical production of tilapia (*Sarotherodon aurea*) and tomatoes (*Lycopersicon esculentum*) in a small-scale recirculating water system. *Aquaculture*, 41: 271–283.

15. Milliken, S. et al. 2022. Analysis of aquaponics research methodology. Cited in: “The Forgotten History of Aquaponics.” iAVs.info. Accessed May 2026.

16. Rakocy, J.E. 1989a. A recirculating system for tilapia culture and vegetable hydroponics in the Caribbean. In: *Proceedings, Auburn Symposium on Fisheries and Aquacultures*, Sept. 20–22, 1984. Brown Printing Co., Montgomery, AL.

17. Lennard, W.A. 2020. A descriptive analysis of the replication applied in aquaponic experimental studies. *International Journal of Environment, Agriculture and Biotechnology* (IJEAB), 5(6).

18. Hurlbert, S.H. 1984. Pseudoreplication and the design of ecological field experiments. *Ecological Monographs*, 54(2): 187–211.

19. “Flawed Research in Aquaponics” (May 2026). Controversy analysis. Documents the “house of cards” effect, the Lennard (2004) confound case study, and systemic pseudoreplication. DRAFTS/controversy/.

20. Lennard, W.A. and B.V. Leonard. 2004. A comparison of reciprocating flow versus constant flow in an integrated, gravel bed, aquaponic test system. *Aquaculture International*, 12: 539–553.

21. Maucieri, C., C. Nicoletto, E. van Os, D. Anseeuw, R. Van Havermaet, and R. Junge. 2019. Hydroponic technologies. In: Goddek, S., Joyce, A., Kotzen, B., and Burnell, G.M. (eds.) *Aquaponics Food Production Systems*. Springer.

22. Silva, L., E. Gasca-Leyva, et al. 2017. Evaluation of a semi-intensive aquaponics system, with and without bacterial biofilter in a tropical location. *Sustainability*, 9(4): 592.

23. Yep, B. and Y. Zheng. 2019. Aquaponic trends and challenges — A review. *Journal of Cleaner Production*, 228: 1586–1599.

24. Pryce, R.T. 2025. Challenges and considerations hindering the validation of aquaponics viability. DRAFTS/controversy/.

25. Flett, I.K. 2017. How aquaponics can improve aquaculture and help feed a hungry world. Dissertation. University of Iceland.

26. Nichols, M. 2015. Aquaponics, sustainable solution or suspect system? *ICESC2015: Hydroponics and Aquaponics at the Gold Coast*, 1176: 123–128.

27. Greenfeld, A. et al. 2019. Economically viable aquaponics? Identifying the gap between potential and current uncertainties. *Reviews in Aquaculture*, 11(3): 848–862.

28. Goodman, E.R. 2011. Aquaponics: Community and economic development. Dissertation. Massachusetts Institute of Technology.

29. Goddek, S., B. Delaide, U. Mankasingh, K. Ragnarsdottir, H. Jijakli, and R. Thorarinsdottir. 2015. Challenges of sustainable and commercial aquaponics. *Sustainability*, 7(4): 4199–4224.

30. Palm, H.W. et al. 2019. Coupled aquaponics systems. In: Goddek, S. et al. (eds.) *Aquaponics Food Production Systems*. Springer, pp. 163–199.

31. “Methodological Rigor in Aquaponics Research: An Evaluation of Feed Composition and Plant Tissue Analytics” (May 2026). Comprehensive methodological audit. DRAFTS/controversy/. Documents the Graber & Junge (2009) data laundering case, the 61% pseudoreplication finding, and the systemic failure of nutrient mass balance in aquaponics literature.

32. “A Critical Analysis of Aquaponics: Methodological Invalidity, Engineered Deficiencies, and Industry Malpractice” (May 2026). DRAFTS/controversy/. Documents the Graber & Junge (2009) n=1 pseudoreplication, the net growing area fallacy, and the “house of cards” effect.