Firstly, Q makes a TCP connection with S.
After that, it sends the first request for the webpage. When it is received, it sees that there is also an image which needs to be picked up, so it makes another request to S for the image.
Then it may close the TCP connection.
This is how it is done if we are using HTTP version 1.1( which is always used now). One connection, also called HTTP persistent connection, is enough to make several requests on the same server.
So, the answer is B.
Before 1.1, there used to be non-persistent connections wherein a client had to make and close N connections for N different requests.