Step 1: When the same interaction is confounded with blocks in every replication of the experiment, information on that one interaction is lost completely; this is complete confounding.
Step 2: When different interactions are confounded with blocks in different replications, then for any given interaction there exist some replications where it is not confounded.
Step 3: Information on each effect can then be recovered from the replications in which it was not confounded, so no effect is entirely lost, only partially, because it is estimated using fewer replications than the rest.
Step 4: This scheme, where different effects are confounded across different replications so that overall information is only partially sacrificed on each, is called partial confounding.
Final Answer: (B) Partial confounding.