It means the thread on which the call is invoked is the one that will be used for the entire operation, even it if examines or operates on every element in the collection.
That distinguishes them from the parallel collections which will use potentially all available processors in parallel and thus multiple threads to conclude the collection operation as quickly as possible.