Looking good..
The movement seems about right.. perhaps the "shudder" from changing it's mind as the threat moves away could be dampened a bit more.
Now that there is a means of perceiving depth other than one fish in front of another.. the space between the first clam and the "glass" appears wider than the image would indicate.. This distance would appear to be less than one third of a clam..however several fish can swim between the glass and the first clam and only partially trigger a close response on the first clam and next to nothing on the second clam.. itself only slightly farther from the glass.
The fishes anti collision is perhaps set to keep the fish too far away from the clams. If the clams are set to close at 3 units the fish seem to be set to stay at least 2 units away... unless forced by another fish.
Criticism for what ever it is worth... Nice work on the clams.. we know they proved to be a real challenge.