We mammals haven't been good poisoners. There are a few primitive forms that secrete toxins—the platypus has poison spines, and an unusual insectivore on a few Caribbean islands, Solenodon, has grooved fangs and secretes a salivary toxin, and itty-bitty shrews have toxic saliva—but our class just hasn't had much natural talent for venom. At least, not recently.

New discoveries of some fragmentary fossils in Canada have shown that there were some flourishing species of small, poison-fanged mammals running around in the Palaeocene, 60 million years ago.

Here, for instance, are the broken jaws of Bisonalveus browni.

Incomplete skull and lower jaws (anterior to right); left upper dentition shown in lingual view. LD, left dentary; LLC, lower left canine; LMX, left maxilla; LPMX, left premaxilla; LUC, left upper canine; RD, right dentary. Scale bar, 5 mm.

a, UALVP 43114, left upper third incisor (at left) and canine (at right) in anterior views showing venom delivery groove in the canine. b, Drawing of upper canine in a. c, Sections through upper canine showing cross-dimensions of venom delivery groove along its length. d, UALVP 43115, isolated left upper canine from locality DW-2 in anterior view, showing venom delivery groove. Scale bars in a and d, 1 mm. e, Mammalia, indeterminate. Isolated right lower canine, UALVP 43116 from locality Cochrane 2, showing labial venom delivery groove; the root of this tooth is missing. Scale bar, 5 mm.

The lower jaws are on top (LD and RD), while the upper jaws are down below—that long pointy thing hanging down, LUC, is the left upper canine. Bisonalveus was a small (that jaw is less than an inch long) rodent-like creature with needle-sharp fangs, and from its dentition and size almost certainly scurried about chowing down on small invertebrates.

Looking more closely at those canines (to the right) revealed a surprise: they were grooved.

Several specimens have a gutter running the length of the upper canines. It's not an artifact of damage to these fossils; they've been sectioned and the groove is enamel-lined. The only things similar to it are the fangs of some poisonous snakes. It is distinct in structure from the fangs of Solenodon, and shrews have no dental specializations to promote introduction of toxins into their prey.

What this tells us is that venoms have evolved multiple times in mammals, and that perhaps there was a period in mammalian history where several lineages explored the strategy of using a poisonous bite. And there were species other than Bisonalveus browni that were doing this: figure e to the right is a canine from an unidentified species with the same attributes, but it is about a half-inch long.

It's actually a little surprising that there aren't more venomous mammals running around. The toxins seem relatively easy to evolve—they are often modified forms of digestive enzymes already secreted in the saliva. Modern shrews, for instance, have saliva loaded with kallikrein-like serine proteases (enzymes that break down proteins) that can paralyze small mammals and frogs, and can cause painful local inflammation in larger animals like us. We don't have a specific explanation for the rarity of poisonous mammals, but I'm relieved—shooing mice out of the house would be much more hazardous if they had tiny ferocious fangs dripping with venom!

Fox RC, Scott CS (2005) First evidence of a venom delivery apparatus in extinct mammals. Nature 435:1091-1093.