Although many journals have done a good job of putting their archives online, it does indeed get pretty difficult to find an electronic copy of any but the most prominent publications if you go back earlier than around the 1970s.
Fortunately, there are still people who know how to deal with this, and whose job it is to help with issues like this: university librarians. When I've faced problems of this sort, working with librarians has been the way to solve it. They often are able to locate a physical copy within their own collection or, if they cannot, often also have relationships with other institutions' libraries who can loan or scan materials from their own collections.
I would also suggest, however, that you may also wish to consider whether you really need to go back this far. Go back the better part of a century, and you will often find that notations and vocabulary have changed to the point where it becomes difficult to understand the paper. Moreover, if the phenomenon has not been entirely neglected, there will often be later reviews or textbooks that present it much better and more comprehensively. Unless you are specifically interested in the historical circumstances of the original work or the perspective of the particular investigator, such later references may turn out to serve your needs much better.