In older papers, one sometimes finds references to sources of funding directly linked to or overseen by military agencies. For example, I have memories of seeing acknowledgments to DARPA funding in a pure mathematics article - much to my surprise, as it didn't seem to have any relevant applications. (However I haven't been able to track down said paper in my personal collection, nor indeed any which contain said references; this might suggest that such support was rare even in the past.)

In more recent papers - at least those I encounter - such attributions are conspicuously absent. Instead those written at American institutions overwhelmingly thank NSF grants for their support. 

**Question.** Is my impression accurate? Namely, did the US military fund pure research with little real-world upshot in the past; and has it become more reticent in recent years? If yes, when and for what reason did this shift occur? 

 - I would be interested in a historical explanation just as much as a personal account.
 - To be sure it is only natural that the defense sector still has a keen interest in certain parts of mathematics, say cryptography or data science. I am specifically interested in (past and present) support for projects where the payoffs seem far-fetched.