Reason number one is just that in general ESA didn't have a lot of probe money to go around in the '80s and early '90s. Thus, they didn't launch a lot of missions in the period that weren't some kind of joint collaboration. The joint missions tended to be ambitious, and thus massive--too heavy for Ariane. Given that the partners usually included the US, with a large stable of medium-heavy lifters, there was no reason to reign this in--better to launch a more ambitious mission than to try to cut it down to launch on an Ariane and only have a fraction of the scientific value of rmuch of the same cost (plus political capital expended in convincing the US to go along with that).
Second, Ariane 1/2/3 didn't really achieve the kind of reliability Ariane has since been known for until...something like the mid to late 1980s. The first 18 launches saw 4 failures. That's a bit riskier than probe missions generally like. Hence, more tech-development-heavy missions like EXOSAT tended to prefer a reliable launch on a more proven US mission--trading cost and political flag waving for reliable launch. By the late 1980s, that reasoning had mostly stopped, and you see more European solo (and even ESA-lead) joint missions using Ariane 4. It's still not a massive number until the 2000s, when ESA probes really started to take off in number. By my count, there were more ESA solo missions between 2000 and 2005 as there were from ESA's foundation to 2000:
Pre-2000: EXOSAT, Giotto, Hipparcos, Infrared Space Observatory, XMM-Newton
2001-2005: PROBA-1, Smart 1, Mars Express, Rosetta, Venus Express