DRAG DROP
Your network contains an Active Directory forest named corp.contoso.com. All servers run
Windows Server 2012.
The network has a perimeter network that contains servers that are accessed from the
Internet by using the contoso.com namespace.
The network contains four DNS servers. The servers are configured as shown in the
following table.
All of the client computers on the perimeter network use Server1 and Server2 for name
resolution.
You plan to add DNS servers to the corp.contoso.com domain.
You need to ensure that the client computers automatically use the additional name servers.
The solution must ensure that only computers on the perimeter network can resolve names
in the corp.contoso.com domain.
Which DNS configuration should you implement on Server1 and Server2?
To answer, drag the appropriate DNS configuration to the correct location in the answer
area. Each DNS configuration may be used once, more than once, or not at all. Additionally,
you may need to drag the split bar between panes or scroll to view content.
Answer: 
Explanation:
Correct: child domain use stubzone
Again I agree with Sjoerd that the correct answer is a stub zone but I disagree with his reasoning. Where do we read that contoso.com is a child of corp.contoso.com?
Stub zones can also be used in a scenario where there is no relationship between 2 domains. What is important in the question is the fact that clients must automatically use the added DNS servers and that is why a stub zone is the best alternative because the DNS servers hosting stub zone (server 1 and 2) regularly update their records by querying the master servers (DC1 and 2) which authoritative.
100% right..