There are a few problems or concerns about the code you've posted.
This for loop body will execute exactly once. It will not execute once per element of com (unless com were a scalar, which it's not.) In addition, since com is a matrix length may not do what you expect. If you want this to run once per row of com use height instead of length or use size(com, 1). If you want this to run once per element use numel.
Since i is a scalar, load(i) refers to exactly one element of the variable.
Both com(i, 1) and com(i, 2) are also scalar, but t is not. Therefore the expression on the right side of the equals sign has the same number of elements as t does.
Can you fit (for example) a dozen eggs into a single cup of an egg carton (without scrambling them)? That's what your code is trying to do. Instead, make the left side of the equals sign refer to the same number of elements as the right, perhaps (since t is a row vector) making the left side a matrix and assigning to a row of that matrix like the following: