A question I came across with an answer i didn't expect

Two peer hosts are running applications that exchange
data using UDP. During the current session, a datagram
fails to arrive at the destination. What is true
concerning the retransmission of the datagram?

1. Datagram retransmission occurs when the
retransmission timer expires in the source host.

2. Datagram retransmission occurs when the retransmission
timer expires in the destination host.

3. Datagram retransmission is controlled by the
application layer protocol.

4. Datagram retransmission involves only the data with
sequence numbers equal to or higher than the sequence
number of the current datagram.

==

The answer supplied is 1, although i'm positive it's 3 since UDP doesn't deal with retransmission in any way as far as i know, and it is written that it's the application layer's job to deal with retransmission when UDP is involved. I would like your thoughts on this, thanks alot by the way, there was also a question that goes like this:

Which of the following subnet masks could be used when
subnetting a Class B IP address? (Choose two.)

1. 255.0.0.0

2. 255.255.0.0

3. 255.192.255.0

4. 255.224.0.0

5. 255.255.252.0

6. 255.255.255.128

==

as far as i can tell the answers are 2, 5 and 6. But the cisco exam says the answers are only 2 and 5, and i for the life of me can't see why 6 wouldn't be valid, if it were a Class C address then yeah it would be a subnet zero situation which they don't qualify as valid, but here they give the net id an extra 9 bits so what the heck are they on about?

Lots of love, in a purely professional sense,
Perdition