Sculpin identification

Read online that a way to distinguish two common Eastern U.S. freshwater sculpins were pelvic fins, Slimy sculpins have 3 rays while Mottled sculpins have 4. Can anyone who has identified or researched these sculpins confirm this, or any other way to identify the two sculpins? I looked at global Slimy sculpin sightings and saw many weren’t RG, so thought that if someone could find a way to distinguish the two, it could help amateur naturalists be able to provide a clear identification for sculpin sightings.

1 Like

Not a fish person, but I found this paper which seems to answer your question:

“Where several species of sculpin occur, they may segregate based on habitat preferences with slimy sculpin occupying colder tributaries and cold-water refugia compared with mottled sculpin found in more intermediate mixing zones (Adams et al. 2015). These two most similar cottids overlap in their distribution in the centre of the continent with discontinuous distribution to the west of Manitoba (Scott and Crossman 1973; updates on fishbase.ca). When there is co-occurrence McAllister (1964) described optimal distinguishing characters, being the number of pelvic fin rays (94% separation; mottled 4, slimy 3), the final ray of the dorsal fin (95% separation; mottled double, slimy single), and the ratio of the length of the caudal peduncle to the postorbital distance (100% separation; mottled <60, slimy >70).”

So using the pelvic fin rays as a character should prove >90% accuracy. The final dorsal fin rays (double in mottled, single in slimy) would seem to be another useful character to distinguish with similar high accuracy. The last character cited (the ratio of the length of the caudal peduncle to the postorbital distance) would provide ideal identification but it won’t be diagnosable in most pics I wouldn’t think.

3 Likes

Thank you for finding this! This provides great ways to distinguish the two, and can serve as a helpful resource for newer and inexperienced naturalists.

1 Like

Why wouldn’t it be diagnosable? The postorbital distance is the distance from the back of the eye to the outermost point on the gill cover. The caudal peduncle is the narrow part of the body where the tail fin attaches. So if you get a good side view, you should be able to compare the proportions of these distances. There are plenty of online diagrams of fish morphometrics to guide you in understanding these measurements.

I didn’t say it wouldn’t be diagnosible from pics, just not in most pics. It would require pics with scale objects in them to allow for precise measurements, and ideally two pics, one of each length/distance. So realistically the observer would really have to have taken the pics specifically for these measurements.

If you just want to ID a fish that you’ve personally caught, have in hand, and you’ve got a measuring tape or calipers, this won’t be too hard. But if you want to ID existing sculpin observations on iNat, it’s likely not going to be too useful. The other two characters cited as much more likely to be visible and diagnosible in iNat observations.

1 Like

I disagree. The character is based on the ratio of the lengths, not on the lengths themselves. So, if the picture is a good side view, I can measure the apparent lengths on the screen and calculate their ratio.

I would disagree (conditionally). I’ve done a fair amount of measuring work with images for morphometrics, and, especially for ratios, small differences in measurements can make a big difference in the final outcome. With pictures, there are issues with curvature of specimens, distance from lens, etc. that can affect the outcome.

It would certainly be possible to test whether or not photos are generally accurate enough for this type of identification or not, but one would need photos of specimens with known ID (or better, known ratio measurements) to see what the error would be. If the method can be validated, great. But I wouldn’t make an assumption that the method developed with manual measurements of specimens in a museum is accurate for IDs based on measurements from photos without explicitly testing it first.