Archaic terminology aside, old (pre-1950) scientific texts are surprisingly readable. Perhaps this is partly due to translation, since many early works come to English by way of Latin, French and German, but I suspect that it has just as much to do with the broad audience of the writings.
An early scientist was not a chemist or a physicist or a geologist or a biologist, but a chemist and a physicist and a geologist and a biologist (at least a little bit). In the modern era, scientists have hunkered down into very specific niches. We don't expect someone to be a chemist and a biologist; we praise them for their interdisciplinary-ness and call them biochemists (or chemical biologists, if you'd rather). We forget how to communicate to those outside our niches. I think this has a lot to do with the masses of impenetrable prose published each week, month and year.
I think there are a few other major factors, too; namely the increased pressure to publish more things more often, which could detract from the quality of individual papers; the increased diversity of backgrounds in science, which likely includes a greater population of people who did not study rhetoric than in Dalton's time; and the affectation of intelligence by using fancy vocabulary as a proxy for actual understanding.