In case this is helpful to anyone (I sat down for 20 min not understanding why B is wrong):

B is wrong based on the rules the stimulus presents (and is not reliant on the "independently owned" condition - this is the part that distracted and confused me). All we know is that if a store sells tropical fish but does not sell exotic birds, it must sell gerbils - TF + /EB → G (→ /I).

What answer choice B does is introduce a new rule for which we have no information. It affirms TF but introduces the co-sufficient condition as pet stores that also sell EB. B's chain reads :

TF + EB → /G. As this is different from the stimulus' rule, we DO NOT know what such set of sufficient conditions yield.

However, on a deeper level, B mistakes sufficient and necessary conditions regarding Exotic Birds. For sake of explanation, I will isolate the Exotic Bird (EB) co-sufficient in the stimulus' rule (the relationship between both conditions is important but this is just for explanation).

Original Chain: /EB → G

Correct Contrapositive: /G → EB

B: EB → /G (B Contrapositive: G → /EB)

In a correct contrapositive chain "EB" would belong in the necessary condition spot. Under B's new rule, Exotic Birds is in the sufficient condition again. Moreover, when compared to the original rule's contrapositive, "/G" belongs in the sufficient condition and not the necessary as answer choice B presents.

I kind of thought of this question as the chemist critiquing the marketing/commercial of this weed killer.

The chemist concludes (simplified) that the data regarding this weed killer is misleading because its effectiveness is dependent on individual soil conditions. Could effectiveness differences be the intention of the weed killer developers? The chemist seems to say no (or else he would not be critiquing the case-by-case soil difference). Hence, the gap in the reasoning is what data is the chemist comparing the weed killer's effectiveness to? I predicted that there must be something wrong about the way the weedkiller was advertised/what promises does it make?

B) Fills in this blank as it says that the weed killer's developers tested it in perfect, unrealistic conditions. Thus, it fills in the chemist's assumption that the actual effectiveness of the weed killer does not match the intended effectiveness because of a data flaw. Therefore, the developer's data cannot be correct.