Hi Ron. In a low dam, there is generally no need to have more than 95% or so of Standard Proctor density, and it is probably going to be constructed with smaller equipment, not Cat 825s. Therefore, there is really no point in using the modified, which is more work. (In a high dam, there are obviously advantages to greater density, but there is generally no need to go to Modified, although occasionally it's done.)
Here's where the more pervious part comes from:
If you compact a fine-grained soil to a given density much dry of optimum, you get get a little different structure, and the individual clods that make up the loose fill prior to compaction don't get smooshed (technical terminology) together quite as well by the sheep's feet. Hence, density within the lift is less uniform, and there is higher permeability (primarily in the horizontal direction). Even if you get a higher density (averaged within the lift) with a lower water content, the soil structure can still be the dominant factor. UNLESS, the dam is high, in which case wetting and the overburden pressure in the lower part of the fill could destroy all of the structure and it won't matter.
Ordinarily, the higher permeability is not a make-or-break for a dam, because it is not all that much higher.
The brittleness is a function of water content and density. If you put in a sample at some specific density (say, 90% of Mod, or 95% of Std, which are comparable), the drier it is, the stiffer and more brittle and prone to cracking it is until it gets saturated (which might not occur until the water is already flowing through transverse settlement cracks). The brittleness is observable in the lab or with a mattock on the fill. Hence, 90% of Mod at Mod OMC would be more brittle than 95% of Std at Std OMC. Also, wetter fill is easier to compact against irregularities in the foundation, structures within the dam, etc.
I have assumed that the OP is referring here to core material, not granular material, because he is talking about sampling to measure density and using the Proctor test. If he is using Proctor for granular material, he may have the wrong test, and he wouldn't be able to get samples that hold together enough for density tests. Proctor is generally limited to stuff with >15% fines, although with very fine sand (Florida sugar sand, for example), it can work OK with less.
DRG