The following example appears in the MATLAB tutorial:
X = [16 2 13;
5 11 8;
9 7 12;
4 14 1]
Using a single subscript deletes a single element, or sequence of elements, and reshapes the remaining elements into a row vector. So:
X(2:2:10) = []
results in:
X = [16 9 2 7 13 12 1]
Mysteriously, the entire 2nd row and the first two elements in the 4th row have been deleted, but I can't see the correspondence between the position of the deleted elements and the index vector 2:2:10. Can someone please explain?
The example you gave shows linear indexing. When you have a multidimensional array and you give it a single scalar or vector, it indexes along each column from top to bottom and left to right. Here's an example of indexing into each dimension:
mat = [1 4 7; ...
2 5 8; ...
3 6 9];
submat = mat(1:2, 1:2);
submat will contain the top left corner of the matrix: [1 4; 2 5]. This is because the first 1:2 in the subindex accesses the first dimension (rows) and the second 1:2 accesses the second dimension (columns), extracting a 2-by-2 square. If you don't supply an index for each dimension, separated by commas, but instead just one index, MATLAB will index into the matrix as though it were one big column vector:
submat = mat(3, 3); % "Normal" indexing: extracts element "9"
submat = mat(9); % Linear indexing: also extracts element "9"
submat = mat([1 5 6]); % Extracts elements "1", "5", and "6"
See the MATLAB documentation for more detail.