Can Neocaridina live with fish and other tankmates?
Evidence review: August 12, 2026
Some shrimp and fish can remain in the same aquarium. That observation is not the same as proving the fish is safe, that newborn shrimp recruit into the colony, or that the shrimp behave as they would without the fish. No controlled N. davidi species-pair trial found in this review followed named aquarium fish long enough to measure adult retention and juvenile recruitment.
The short version: decide what success means first. For a line-breeding or growing colony, use a dedicated shrimp tank. For a community display, treat coexistence as a documented tradeoff, provide an exit tank and never turn one peaceful observation into a species guarantee.
Three goals that need different answers
| Goal | Outcome that matters | Lowest-uncertainty design |
|---|---|---|
| Community display | Known adults remain observable and uninjured | A reversible trial with records and a prepared removal destination |
| Growing colony | Hatched cohorts repeatedly reach later visible stages | A dedicated shrimp tank removes tankmate predation and competition |
| Line breeding or research | Parentage, grade, survival and recruitment remain interpretable | No fish or unrecorded mobile tankmates in the production tank |
A tank can meet the display goal and fail the recruitment goal. Seeing the same adults for weeks does not answer what happened to newly hatched shrimp.
What direct Neocaridina evidence shows
Fish cues can change behaviour without a fish touching a shrimp
A 2026 study tested captive-bred N. davidi in five-minute arenas. The predator treatment used water that had housed green swordtails (Xiphophorus hellerii), not a live fish in the arena. On average, shrimp shifted toward a multivariate behaviour the authors interpreted as shyer, but the mean change across individual traits was modest at 0.237 standard-deviation units.1
The full study used 1,191 offspring from 75 full-sibling families nested within 37 paternal half-sibling groups, all descended from one pet-trade-source colony. Eighty shrimp completed six alternating assays and 1,111 completed one treatment-balanced assay. Full siblings shared a rearing tank, which limits how cleanly family, environment, parental effects and additive genetics can be separated.
That experiment establishes a response to waterborne cues under its test conditions. It did not measure capture, injury, survival, feeding, reproduction or long-term coexistence, and the authors deliberately used a novel predator. It therefore cannot label swordtails, or any other fish, safe or unsafe in an aquarium.
Shelter use is normal even without fish
In a 72-hour study with no fish present, N. davidi were recorded using shelter in 82.1% of observations on average. Shelter use was higher by day than by night. Of the four tested structures, Vesicularia moss was used most, followed by wood, while rock was used least.2 The experiment included males, non-ovigerous females and ovigerous females and found differences in activity by sex and reproductive state.
This supports offering structurally useful shelter. It does not show that moss is predator-proof, calculate how many juveniles it saves, or turn a fish pairing into a safe one. Hiding also cannot be assigned to fear from appearance alone because shelter use and daily activity vary without fish.
One individual's behaviour is not a species rule
Repeated open-field and food-shelter tests in 48 captive red cherry shrimp found repeatable among-individual differences in several behaviours.3 The study included no predator and does not predict which shrimp will survive a community tank. It does show why one unusually bold or concealed animal cannot stand in for every shrimp, line or life stage.
Why there is no safe-fish list here
Labels such as peaceful, nano, algae eater, bottom-dweller and community fish do not measure a shrimp outcome. A useful screen needs the exact species, individual size and history, feeding mode, water requirements, occupied tank zone, activity period and the smallest shrimp stage that matters to the keeper.
| Question | Why it matters | What it cannot promise |
|---|---|---|
| Do the measured water and temperature needs overlap? | Both species must be maintained in defensible conditions | Overlapping water does not prevent predation |
| Can the tankmate enter the same surfaces and openings? | Encounter opportunity changes with body size, life stage and structure | A small mouth or different tank zone is not a guarantee |
| How does the individual respond to moving invertebrates? | Individuals can differ within a named species | Ignoring one adult once does not predict later encounters |
| Is the goal adult display or recruitment? | Newly hatched shrimp and adults are different exposure targets | Adult retention does not establish juvenile survival |
| What extra food, waste, oxygen demand and disturbance are added? | Tankmates change more than predation risk | Gallons or filtration alone do not calculate the combined capacity |
| Where will either animal go if the trial fails? | Removal must be possible before injury or unexplained loss accumulates | A divider or net found later is not an exit plan |
A reversible community-display protocol
Our practice
Do not run this experiment in a line-breeding, quarantine or irreplaceable colony. If losing shrimp is unacceptable, the decision is already made: keep the animals separate.
- Write the exact tankmate species, number, approximate sizes and known feeding behaviour. Verify that the operating water is defensible for every species.
- Define success before introduction: adult display, adult retention, or repeated juvenile recruitment. Select the observation window and count method for that outcome.
- Record a pre-introduction baseline using the same viewing times and tank zones: adult range, berried females, hatch observations, visible juvenile stages, feeding response, distribution, deaths and missing animals.
- Provide complex, multi-scale surfaces before the trial. Confirm that the smallest retained shrimp stage can enter some spaces the tankmate cannot. Treat this as risk reduction, not proof of protection.
- Prepare and commission a separate destination before introduction. Add only one tankmate species or one defined change at a time.
- Observe across daytime, lights-off and feeding periods. Record pursuit, ingestion, injury, displacement from food, changed distribution and the same population outcomes used in the baseline.
- End the trial after observed pursuit or ingestion, injury, repeated unexplained loss, inability to feed, or deterioration in the outcome chosen at the start. Missing shrimp remain missing, not confirmed predation, but they still weaken a retention claim.
How to read common observations
| Observation | What it establishes | What remains unknown |
|---|---|---|
| The fish ignored one adult during observation | No pursuit was observed in that window | Night activity, future encounters and juvenile outcome |
| Adults are still visible after the chosen interval | Those visible adults remained in the system | Total retention, injury, stress and recruitment |
| One newly hatched shrimp is seen | At least one hatchling was visible at that time | Hatch count and survival to a later stage |
| Later juveniles repeatedly appear | Some recruitment occurred under those recorded conditions | Total losses and what the outcome would have been without tankmates |
| Shrimp hide more after introduction | Distribution or visibility changed | Cause, welfare consequence and whether the population declined |
What about snails and other invertebrates?
Invertebrate is not a compatibility category. Identify the exact animal and apply the same screen for water, predatory feeding, body-size access, food competition, disease movement and exit planning. Quarantine still matters. Mixing colour lines of N. davidi also creates a genetics decision even when direct aggression is not the concern.
Do not add a fish or another animal to control an unidentified worm or polyp in a breeding colony. Consumption would not prove eradication or safety for juvenile shrimp. Identify and contain the organism first.
What the evidence does not prove
- That any named aquarium fish is universally shrimp-safe
- That a small mouth makes a fish safe for newly hatched shrimp
- That peaceful, herbivorous or algae-eating labels predict coexistence
- That moss guarantees juvenile survival
- That adult survival proves a colony can grow
- That no observed attack means no predation occurred
- That one keeper's successful pairing predicts another tank
- That snails or other invertebrates are automatically harmless
Primary sources
- Wilson, Rickward and Santostefano, 2026. Quantitative genetics of shy-bold behaviour and plastic response to novel predator cues in the cherry shrimp, Neocaridina davidi. Journal of Evolutionary Biology, voag042.
- Carvalho-Batista et al., 2023. Shelter preference and variation in the daily activity pattern of the ornamental shrimp Neocaridina davidi. Nauplius 31: e2023018.
- Rickward, Santostefano and Wilson, 2024. Among-individual behavioural variation in the ornamental red cherry shrimp, Neocaridina heteropoda. Ecology and Evolution 14(2): e11049.
Measure the combined system's operating capacity.
Use the source-specific quarantine and release guide.
Use the health-triage guide for abnormal observations.
Interpret hiding and activity without turning them into a compatibility verdict.