A video filmed at 300 fps of an axolotl trying to snatch up a food pellet. (Unfortunately, it fails both times to swallow the pellet.) Playback is slowed down to about 10x, and then 50x the actual speed of the feeding reflex. Filmed with a Casio Exilim EX-F1. Video quality is certainly not the best, but the video is still interesting, I think.

Embedded YouTube video:

Timing stats

When crudely measuring the delays between events in 4 video sequences, I find the following numbers:

delay 1: Time from impact of food pellet to visible reaction of axolotl
delay_1 = [30   37   43   63];
mean_1 = 43 ms
std_1  = 14 ms


delay 2: Time from visible reaction to fully opened mouth
delay_2 =  [17   23   27   40];
mean_2 = 27 ms
std_2 = 10 ms

A reaction time of about 40 ms isn't too shabby; and the mouth snaps open wickedly fast within about 30 ms. The variability in timing probably arise from the axolotls sitting at different depths that I didn't control for. Presumably, detection of the food pellet happens through mechanosensors of the lateral line organ that detect the slight pressure changes that happen when a pellet hits the water. Measuring the delays in a more controlled way would be a nice science projects for kids…