Learn · Articles

KALARVO: black soldier fly larvae on Kainuu fish side streams – what three trial rounds showed

What three rearing rounds of black soldier fly larvae on Kainuu fish side streams showed, from recipe shares to the route to commercial use.

01 · Summary

Summary

Three rearing rounds, 70 recipes and laboratory analyses of the larvae and the frass. This summary sets out what worked for each fish side stream, what the lab found and what the route to commercial use looks like.

  • All three fish side streams became larvae. Across three rounds (28 April to 18 June 2026) we reared 70 recipes. There was no single best raw material or recipe: each side stream has its own role.
  • Fish recipes beat the fish-free control. In round 1, 31 of 35 successful fish recipes (89 %) outperformed it; in round 2, all 15; in round 3, all seven successful fish recipes.
  • The lake-fish side stream was the most reliable. Not one of its batches failed. Its best result was 1.38 kg of fresh larvae per tray at 145 mg per larva. Heads and bones need effective crushing.
  • Moisture in the mix mattered more than anything else. The same water setting that suited lake fish and filleting side stream drowned rainbow trout gut mass: in the final round, five of six gut-mass batches failed wet.
  • The fish produced two different larvae. Larvae reared on the lake-fish side stream were lean and protein-dense (48 % protein and 25 % fat in dry matter); larvae on the trout filleting side stream were fatty (41 % and 41 %). Both contained measurable EPA and DHA.
  • Mercury follows the fish into the larvae. Mercury in the lake-fish larvae was about 2.5 times the general limit for feed materials. The filleting-side-stream larvae were well below every heavy-metal limit.
  • The frass is rich in nutrients and needs hygienisation. It held 4.0–4.8 % nitrogen in dry matter and little heavy metal; as a fertiliser it must first be heat-treated.
  • Technically yes, legally not yet. Raw fish side streams are not currently a permitted rearing substrate for insects grown for feed or pet food. Commercial use needs a route confirmed in writing.

02 · Background

Background and aims

Black soldier fly (BSF) larvae eat organic material and turn it into two products. Larval biomass is a protein-rich raw material for feed and pet food. Frass, the rearing residue, is a mix of larval droppings and uneaten material; once treated, it can be used as a fertiliser or soil improver.

Fish processing in Finland's Kainuu region produces side streams that have to be handled anyway. In the KALARVO project, Kajaani University of Applied Sciences (KAMK) wanted to know whether larvae could upgrade them, how each side stream is best used, and under which conditions this could make commercial sense. The trials answered three questions, separately for each raw material:

  1. How should the side stream be used, and what fish share works best?
  2. What kind of larvae and frass does it produce, and what are they suitable for?
  3. Under which conditions could the use be viable?

Three fish raw materials were tested:

  • Rainbow trout gut mass: the viscera from gutting farmed rainbow trout. It was tested in two forms: slowly thawed with the surface oil skimmed off, and acidified.
  • Rainbow trout filleting side stream: heads, frames, fins and skin, with the flesh left on them.
  • Lake-fish side stream: gut mass and filleting side stream from mixed wild lake fish such as bream, pikeperch, pike, perch and vendace, delivered combined. It is the side streams of these fish, not whole fish.

Besides fish, the recipes used supporting ingredients: potato (unsaleable raw potato or potato peeling side stream), brewer's spent grain to give structure and bind moisture, pre-moistened wheat bran to adjust moisture and, in some recipes, 1 % garlic.

03 · Method

How the trials were run

The trials ran in a 20-foot Manna MIND rearing container. The MIND system controlled temperature, humidity and ventilation automatically and logged them, together with carbon dioxide, continuously to the cloud. Relative humidity held at about 70 % in every round (72.3 % in round 1, 69.8 % in round 2). Every recipe had the same climate, so differences in results came from the recipe. Manna Insect was responsible for the trial design, the measurements, the analysis and the reporting. More on the container: the MIND container article.

Each round followed the same chain: frozen fish was thawed and mixed by weight with the supporting ingredients into 60 × 40 cm trays, 3–5 replicate trays per recipe. The trays warmed overnight to about 30 °C, then each received 10,000 five-day-old larvae with all the food at once. The target cycle was seven days. At harvest the larvae were sieved from the frass and weighed, and larval size was measured from a 100-larva sample.

  • Round 1 (screening, 39 recipes). A broad screen of all three fish raw materials. Three batches with a lot of fish and potato but no brewer's spent grain stayed a sticky mass; spent grain proved the best ingredient for structure and moisture binding. The largest larvae (183 mg) grew on acidified gut mass.
  • Round 2 (expansion, 17 recipes). The mix was made drier and gaps were filled. No batch failed, and acidified gut mass performed the same as the oil-skimmed form. Batches finished early because the mix dried out.
  • Round 3 (final, 14 recipes). The best and the new promising recipes plus a fish-free control. The lake-fish side stream gave the strongest result of the project, while five gut-mass recipes failed wet: the larvae left the over-wet trays before they started eating. Larvae and frass from the successful recipes went to the laboratory.

04 · Results

Results by raw material

Fish against a fish-free feed. Every round had a fish-free control. In rounds 1 and 2 it was a wheat-bran-based plant feed with no animal protein at all; in round 3 it was a fish-free mix of potato, brewer's spent grain and wheat bran. In round 1, 31 of 35 successful fish recipes beat the control, in round 2 all 15, and in round 3 all seven successful fish recipes, on both larval size and larval mass.

All three fish side streams turned organic matter into larvae about equally well: bioconversion was 16–18 % for each. The differences lay in reliability, larval size and how easy the material is to handle.

Lake-fish side stream: the base raw material. The lake-fish side stream was the most reliable material in the trials. None of its batches failed in three rounds, it tolerated moisture swings best, and the harvest came out clean. The best final-round recipe with raw potato produced 1.38 kg of fresh larvae per tray at 145 mg per larva. The same recipe with potato peeling side stream gave 1.21 kg and 130 mg, so peelings are a practical alternative. Heads, frames and bones do not break down, so the material must be crushed before mixing.

Rainbow trout gut mass: big larvae, narrow moisture window. Gut mass has the highest ceiling for larval size. In round 1, acidified gut mass with potato produced larvae averaging 183 mg per batch. It is also the most moisture-sensitive material: in round 1 the larvae escaped from over-wet trays, and in round 3 five of six gut-mass batches drowned. Raising the fish share did not grow bigger larvae, it only added water and risk. The acidified form worked as well as the oil-skimmed one, which saves the skimming work, in the dry range. Gut mass is uniform and needs no crushing.

Rainbow trout filleting side stream: a co-substrate. The filleting side stream worked at every share and tolerated moisture, but heads, frames and skin did not break down; they caused clumping, an ammonia smell and uneaten food. The cleanest results came at about 20 % fish. Fine crushing is the key to using it.

Raw material

  1. Lake-fish side stream

    Recommended fish share
    30–45 %
    Best recipe (total shares of the mix)
    Round 3: lake-fish side stream 27 %, raw potato 27 %, wheat bran (dry) 18 %, added water 27 %*
    Result
    1.38 kg/tray, 145 mg/larva
    Role
    Base raw material; needs crushing and mercury screening

  2. Rainbow trout gut mass

    Recommended fish share
    10–30 %
    Best recipe (total shares of the mix)
    Round 1: acidified gut mass 30 %, potato 15 %, pre-moistened wheat bran 55 %**
    Result
    183 mg/larva (batch average)
    Role
    Co-substrate; dry and at a low share

  3. Rainbow trout filleting side stream

    Recommended fish share
    20–40 %
    Best recipe (total shares of the mix)
    Round 2: filleting side stream 20 %, potato 15 %, brewer's spent grain 20 %, pre-moistened wheat bran 45 %**
    Result
    1.00 kg/tray, 106 mg/larva
    Role
    Co-substrate; needs fine crushing

* Round 3 shares include the water added at loading. Shares are rounded.

** The wheat bran share includes the water used to pre-moisten it.

Moisture was the single most important factor. The wheat bran pre-mix was tuned every round: first water to bran 2:1 (too wet), then 1:1 (too dry), and finally 1.5:1 plus water added at loading. Recipes with a high 40–45 % gut-mass share, which worked in the drier round 2, drowned in round 3. Moisture tolerance is a property of the raw material, so each fish side stream needs its own water setting.

Air humidity did not decide mix moisture. The same humidity of about 70 % sat alongside both over-dry and drowned mixes. The recipe decides how wet the mix is.

Food ran out before day seven. Working batches finished earlier each round, around day 8, 6 and 5, because the mix dried out. When food lasted, larvae kept growing: two trays reared to day 8 produced the largest individual larvae of the project, 195 and 187 mg. A seven-day cycle is a good production target when moisture and food last until day seven.

The results are in line with published research. Bioconversion was in the same range as the literature, and the best lake-fish result (145 mg) sits inside the 100–180 mg window published as optimal for fish substrates.

05 · Composition

Larvae composition

Larvae from the successful fish recipes of round 3 were analysed in an accredited laboratory (ISO/IEC 17025).

In dry matter

  1. Crude protein

    Lake-fish side stream
    48.2 %
    Trout filleting side stream
    40.7 %

  2. Crude fat

    Lake-fish side stream
    24.7 %
    Trout filleting side stream
    41.0 %

  3. Ash

    Lake-fish side stream
    8.5 %
    Trout filleting side stream
    5.8 %
  • Two different larvae. Larvae from the lake-fish side stream were lean and protein-dense. Larvae from the filleting side stream were fatty, because the fat in the raw material carried over into the larvae. The bone in the lake-fish material showed up as more ash, calcium and phosphorus.
  • Gut-mass larvae sit in between: about 44 % protein and 32 % fat in dry matter (estimate).
  • The same protein quality. The amino acid profile was practically identical in all three groups. The fish changed how much protein there was, not its quality. Lysine was good; methionine and cystine were, as usual for BSF larvae, the scarcest group.
  • Omega-3 is measurable. The larvae contained EPA and DHA together at 0.41 g (lake fish) and 0.60 g (filleting side stream) per 100 g of fresh larvae: a real but moderate enrichment, not a fish-oil replacement.

06 · Safety

Mercury

Mercury is the main safety finding of the trials. Larvae reared on the lake-fish side stream contained 0.097 mg/kg of mercury in fresh weight. Feed-material limits are set at 12 % moisture, where this equals about 0.25 mg/kg, about 2.5 times the general limit of 0.1 mg/kg for feed materials.

Larvae reared on the trout filleting side stream had 16 times less mercury, about 13 % of the limit. Cadmium, lead and arsenic stayed below 10 % of their limits in both groups.

Mercury passes from the fish into the larvae. Using the lake-fish side stream for feed or pet food therefore needs mercury screening of each raw-material batch and its species mix.

07 · Frass

Frass

Round 3 frass was analysed for nutrients from all three fish groups, and for heavy metals and hygiene from the lake-fish group.

In dry matter

  1. Total nitrogen (N)

    Lake-fish side stream
    4.0 %
    Trout filleting side stream
    4.8 %
    Trout gut mass
    4.6 %

  2. Phosphorus (P)

    Lake-fish side stream
    1.98 %
    Trout filleting side stream
    0.91 %
    Trout gut mass
    0.70 %

  3. Potassium (K)

    Lake-fish side stream
    1.16 %
    Trout filleting side stream
    0.85 %
    Trout gut mass
    1.07 %
  • Nitrogen-rich. All three frass samples held 4.0–4.8 % nitrogen in dry matter, above the roughly 3 % typical of commercial BSF frass. Phosphorus followed the bone content of the raw material and was highest in the lake-fish frass.
  • Low in heavy metals. Mercury in the lake-fish frass was 0.15 mg/kg of dry matter, about half the level in the larvae from the same group and about 15 % of the organic fertiliser limit. Cadmium was about 12 % of its limit, lead and arsenic below 1 %. The mercury risk sits with the larvae, not the frass.
  • Hygienisation needed. Salmonella was not detected in a 25 g sample, but E. coli was above 2,400 MPN/g. Frass used as a fertiliser is held at 70 °C or more for 60 minutes in an approved process, after which E. coli may be at most 1,000 cfu/g.

Once hygienised, the frass is a useful nitrogen-rich by-product.

08 · Regulation

Regulation and route to use

The trial result is a technical result, not a commercial permit. The current guidance of the Finnish Food Authority (Ruokavirasto) and the list of permitted animal-derived rearing substrates in Regulation (EU) 2017/893 include fishmeal, but not raw fish or the side streams of gutting and filleting. The raw, frozen or acidified fish side streams tested are therefore not currently a permitted rearing substrate for insects grown for feed or pet food.

KALARVO showed that bioconversion works. Before commercial use, the exact product, raw-material and facility chain is confirmed in writing with the Finnish Food Authority.

The usability of each side stream in principle:

  • The lake-fish side stream can be valuable to a grower if the material is consistent, ready to process and delivered nearby.
  • The rainbow trout gut mass has conditional value. Wet or variable material may require the grower to be paid for receiving and handling it.
  • The trout filleting side stream depends on pre-processing, and on who is responsible for crushing and logistics.

Fish brings real biological value to the larvae; collection, stabilisation, crushing and moisture control decide whether a side stream is a raw material or a handling problem for the grower.

09 · Next steps

Next steps

The recipe questions have been answered well enough that no new broad screening is needed. A focused follow-up pilot is justified once a permitted use is confirmed. The next phase:

  • A written authority route for the exact product, raw-material and facility chain, or a move to a permitted substrate.
  • One or two confirmation trials per raw material.
  • Fixed moisture windows per raw material, measured at loading.
  • Seven-day moisture for lake fish, to raise both size and yield.
  • A practical crushing solution for the filleting and lake-fish side streams.
  • Mercury screening of lake-fish raw-material batches and their species mix.
  • Hygienised frass and its analyses.
  • A local replacement for wheat bran, which is not available in Kainuu.
  • Recipe-specific climate for each winning recipe.
  • Real delivery, cost and use terms for each side stream.

If you have a side stream of your own that you want evaluated for larval rearing, get in touch with Manna Insect.

10 · About the project

About the project

Kajaani University of Applied Sciences (KAMK) procured black soldier fly larvae rearing trials on Kainuu fish side streams as part of the KALARVO project (Processing fish side streams into value products). Manna Insect carried out the trials as the contractor, responsible for the trial design, measurements and analysis. Co-funded by the European Union.

KAMK, Kajaani University of Applied Sciences Co-funded by the European Union

KALARVO project · Kajaani University of Applied Sciences (KAMK) · Co-funded by the European Union

Keep exploring

Related topics and pages

Read next

Next step

Evaluate a side stream for larval rearing

Manna Insect carried out the KALARVO trials and can help evaluate your own side stream for black soldier fly rearing, with a trial design that matches your material.

Get in touch

Back to Articles