Scores explained
Every species gets two separate scores, plus an average of the two. This page explains what goes into each one, where the numbers come from, and how far to trust them.
Venom Hazard
How dangerous the venom itself is: how toxic it is, and how much of it the snake can deliver.
Human Risk
How likely a bite is to happen and to go badly for the person bitten.
Combined
Both scores averaged, for a single sort order. Always read it with the two scores beside it.
Venom Hazard: how dangerous is the venom itself?
Two measurements from the lab: how toxic the venom is per milligram (LD50), and how much venom the snake can produce (venom yield). A snake needs both to be truly dangerous. Very toxic venom in tiny amounts, or huge amounts of mild venom, both score lower than a snake with a lot of potent venom.
LD50: how toxic the venom is
LD50 is the dose of venom that kills half of a group of test animals, usually mice. It is written in milligrams of venom per kilogram of body weight (mg/kg). The lower the number, the more toxic the venom.
When a headline calls a snake “the most venomous in the world”, it almost always means that snake has the lowest LD50.
What 0.025 mg/kg looks like
The inland taipan’s venom has an LD50 of 0.025 mg/kg when injected under the skin of mice (Broad, Sutherland & Coulter, 1979). For a typical 20-gram lab mouse, that works out to half a microgram.
We do not multiply this up to a “human lethal dose”. People are not large mice: body size does not scale toxicity in a straight line, and species respond differently to the same toxins.
How the test works
Researchers give several groups of mice increasing doses of venom and count how many die within a fixed time, commonly 24 or 48 hours. Plotted against dose, the results form an S-shaped curve. The LD50 is the dose where that curve crosses 50%.
The 50% point is calculated with statistical methods such as probit analysis, Spearman–Kärber or Reed–Muench, and a good study reports a confidence interval with it. Groups are often only 5 to 10 mice per dose, so that interval can be wide.
Where the venom goes changes the number
The same venom gives a different LD50 depending on how it is injected. This is the biggest source of confusion when figures from different studies are compared.
| Route | Injected into | What it is used for |
|---|---|---|
| SC | Under the skin | Closest to a real bite. Our preferred route. |
| IV | A vein | The most standardised route; venom reaches the blood at once. Common in antivenom testing. |
| IP | The belly cavity | Historically the most common route because it is easy, but absorption is less predictable. |
| IM | A muscle | Less common; used in some tissue-damage studies. |
Even the liquid the venom is dissolved in matters. In the same 1979 study, inland taipan venom measured 0.025 mg/kg in plain saline and 0.01 mg/kg when diluted with a protein carrier (bovine serum albumin). We use the saline figure and keep the other on record.
All 94 species on one scale
LD50 values span more than four orders of magnitude, from 0.025 to over 700 mg/kg, so they only make sense on a logarithmic scale. Each step on the axis below is ten times the one before. Filled dots were tested under the skin (SC); hollow dots used another route or an unrecorded one.
What LD50 is good for, and what it misses
- One comparable number. It puts very different venoms on a single scale.
- Testing antivenom. WHO guidelines measure antivenom strength by how many lethal doses of venom it neutralises.
- Decades of data. Results going back to the 1970s can still be compared.
- People are not mice. Mouse results do not reliably predict what a bite does to a person.
- How much venom is delivered. LD50 is toxicity per milligram. That is why we add venom yield, below.
- Injury that doesn’t kill. Tissue death, bleeding and lasting disability do not show up. Many viper bites maim far more often than they kill.
- Slow killers. The test favours fast-acting nerve toxins and underrates venoms that kill over days through bleeding or kidney failure.
- Natural variation. Venom changes with a snake’s age, region and diet, but studies often pool venom from several snakes.
- How death happens. A single number says nothing about whether it was paralysis, bleeding or heart failure.
Venom yield: how much venom the snake has
Venom yield is the amount of venom collected when a snake is “milked”: it bites through a membrane stretched over a container, or its venom glands are gently pressed. The venom is freeze-dried and weighed, and the result is given in milligrams of dry venom.
Sources usually give a typical figure and a maximum. We score the maximum. Many older sources give only a range without saying what the lower number means, so the maximum is the more consistent figure to compare.
What yield misses
- Milked is not injected. Snakes control how much venom they use. Many defensive bites inject a fraction of the maximum, and some are “dry”, with no venom at all.
- Big snakes, big yields. Yield rises with body size, so a record from one large captive snake can sit far above what a typical wild snake produces.
- Thin evidence. Most yield figures come from older secondary compilations that don’t say how many snakes were milked. That is why most yields in our data are grade C.
Why both measurements matter
Two snakes with almost the same Venom Hazard, for opposite reasons.
How the two become one score
Worked through for Russell’s viper.
- Pick one LD50 per species
Under-the-skin (SC) tests are preferred. Competing figures stay on record, and a species is flagged when they differ by more than 1.5×.
0.03 mg/kgSC · grade C - Score potency from 0 to 100 on a log scale
100 at the most toxic anchor (0.039 mg/kg), 0 at the least toxic (9.27 mg/kg). The anchors are the 2nd and 98th percentiles of all species, so one extreme value cannot stretch the scale. Anything beyond an anchor is capped.
Potency 1000.03 is past the anchor, so capped - Score quantity from the maximum yield, also on a log scale
From the smallest maximum yield in the dataset (0.7 mg) to the largest (1,102 mg).
250 mg → 79.8 - Average the two
Potency and quantity count equally by default. Both are required: a species missing either gets no Venom Hazard.
Venom Hazard 89.9rank 1 of 78
Human Risk: how likely is a bite to happen and go badly?
Venom alone doesn’t decide how many people a snake harms. Human Risk combines four things about the snake and the people around it. Unlike the lab measurements, these are informed judgements, so treat the score as indicative.
| Factor | The question | Scale | Where it comes from |
|---|---|---|---|
| Encounter | What are your chances of encountering this snake? Very low for sea snakes and the inland taipan. | 1–10 | Author estimate from range, habitat and how common the snake is near people. |
| Bite propensity | When met, how readily does it bite? | 1–5 | Assigned from documented behaviour, using the scale below. |
| Onset | How fast does an untreated bite progress? | 5 levels | Clinical descriptions of envenoming. |
| Antivenom access | Is there an antivenom, and is it available where the snake lives? | 5 levels | Antivenom listings and regional availability. |
Bite propensity
| Level | Behaviour |
|---|---|
| 5 | Strikes readily and repeatedly, or pursues a threat |
| 4 | Bites readily when disturbed, or holds still and strikes when stepped on |
| 3 | Defensive if cornered or provoked; usually warns or retreats first |
| 2 | Shy or docile; bites mainly if handled or stepped on |
| 1 | Rarely bites, even when handled |
Onset: faster scores higher
The faster an untreated bite becomes life-threatening, the less time there is to reach a hospital that holds antivenom. For many people bitten far from care, a venom that acts within an hour leaves no time at all. So the fastest-acting venoms score highest, and the slowest score lowest.
| Untreated onset | Scores |
|---|---|
| Very fast (under 1 h to a few hours) | 90 |
| Fast (hours) | 70 |
| Moderate (6–24 h) | 50 |
| Moderate to slow (hours to days) | 30 |
| Slow (24 h or more) | 10 |
Antivenom access
| Access | Scores |
|---|---|
| Widely available | 10 |
| Regional | 30 |
| Limited | 50 |
| Very limited or unconfirmed | 70 |
| None | 90 |
How the four become one score
Each factor is placed on a 10–90 scale. The lowest level scores 10, not 0, because a category is a band, not “no risk”. The four are then averaged with these weights:
If a factor is unknown, the average uses the others. A species needs at least three of the four for a score, and scores are most comparable between species with the same number of factors.
The combined score
The combined score is Venom Hazard and Human Risk averaged with equal weight. It gives one sort order for people who want a single list.
Use it with care. It blends the two questions we separated, so a middling combined score can hide very potent venom on a snake that people rarely meet, or the reverse. On every page we show it next to its two parts.
How far to trust a number
Published figures are often copied from list to list with the original study lost along the way. So every value on this site carries its source and a grade.
| Grade | What it means | How we use it |
|---|---|---|
| A | A named, peer-reviewed study with its method stated | The gold standard |
| B | An expert database, WHO document, or a study reached through one | Accepted |
| C | A secondary compilation that doesn’t restate the method | Fine for comparison; flagged |
| D | No traceable source | Shown with a warning until a source is found |
A species’ data grade is the worse of its LD50 grade and its yield grade. Wikipedia tables and “most venomous” rankings are not accepted as sources at all.
Why this matters: the hook-nosed sea snake
Three LD50 figures circulate for this species, and they differ eightfold.
We use the 0.164 figure. The other two stay on record, and the species is flagged because its sources disagree. Switching dropped its Venom Hazard from 82 to 69.
Other measures toxinologists use
| ED50 | The dose of antivenom that protects half the animals from a set amount of venom. Used to test antivenoms. |
| MHD / MND | Minimum haemorrhagic or necrotic dose: the smallest dose that causes a set amount of bleeding or tissue death. More relevant for many viper venoms. |
| Lab-dish tests | Clotting, enzyme and cell tests that measure what a venom does without killing animals. |
| Venomics | Identifying every protein in a venom by mass spectrometry. It shows how a venom works, not only how lethal it is. |
Sources
- Broad, A.J., Sutherland, S.K. & Coulter, A.R. (1979). The lethality in mice of dangerous Australian and other snake venom. Toxicon 17: 661–664.
- World Health Organization (2017). Guidelines for the production, control and regulation of snake antivenom immunoglobulins.
- Ernst, C.H. & Zug, G.R. (1996). Snakes in Question. Smithsonian Institution Press. Secondary compilation behind many grade C values.