Okay, I know this is kind of dumb, but it's something I've noticed:
Obviously, we know that x0 and x5 Interstates are generally the longest and most significant. However, I've noticed that when those are used up, it seems preference is then given to x4 and x9 numbers. For example, I-84 and I-94 are relatively long "non-primary" Interstates. And then down in Florida, they chose I-4 over I-2 and I-6. You have important south-north corridors such as I-29, I-39, I-59, etc, when other numbers like 31, 41, 61, etc. go unused.
Am I just looking too deeply into all this, or does it seem somewhat intentional? It does seem that by picking a middle number like 4, it leaves some "wiggle room" for future Interstate corridors that might be located between a x0 and x4.
Now that you mention it, it does seem like there is a slight favoritism towards x4's and x9's. It's not so clear-cut, but it is kind of there.
I'd throw I-74 into that list, as it may ultimately be a pretty long interstate if the Carolina and Illinois-to-Ohio segments are connected. Ditto I-64 given its length. I-24 and I-44 are also fairly prominent and moderately-lengthy interstates. So there might be something to what you say.
This trend favoring 4's and x9's might stem from desire to "space out" the numbers, as numbers ending in 4 and 9 are about intermediate between consecutive x0's and consecutive x5's. But the same would be true of x6's and x1's, so it still doesn't explain why x4's and x9's get preference. Maybe it's something as arbitrary as the personal preferences of members on the original numbering committees?
Maybe if you took all the Interstate, US and State routes in the country, listed them all by their final digit, 4's and 9's are the least used?
I realize these categories may be arbitrary- you could quibble over certain routes I'm sure, but I'd say this is more or less right, so but let's try that
Trans-national (more or less)
10, 20, 40, 70, 80, 90
5, 15, 25, 35, 55, 65, 75, 95
Nationally significant
30
71, 81, 91
73 (the entire I-73 corridor), 93
24, 44, 64, 74, 84(E), 84(W), 94
45, 85
26, 76(E)
57, 77 , 87
8
29, 49 (once complete), 59, 69
Regional connectors
4
12, 22 (once complete), 72 , 82
43, 83
16, 66, 76(W), 86(E), 86(W), 96
17, 27, 37, 97
68, 78, 88(E), 88(W)
19, 39, 79, 89, 99
Not a thing
0, 50, 60
1, 11, 21, 31, 41, 61
2, 32, 42, 52, 62, 92
3, 13, 23, 33, 53, 63
14, 34, 54
6, 36, 46, 56
7, 47, 67
18, 28, 38, 48, 58
9
Let's see- if you count duplicates and assign "Not a thing" a value of 0, "Regional Connector" a 1, "Nationally important" a 2, and "Trans-contiental" a 3 (yes, this is arbitrary- if somebody wants to do this better go for it. Using raw length seems like a bad idea because the number distribution is skewed eastward and I'd argue that length has less to do with importance as you move eastward (a short route in the east can serve as much population as a long one in the west)-it would give too much priority to things like I-83, but it might be better than doing it this way). This gives some sort of weight to frequency of number assignment and importance of routes with that assigned number, which if I'm not too arbitrary (and I may very well be- if somebody wants to spend more time, be my guest. I-91 is obviously not twice as important as I-43- so take with a grain of salt.) should reveal relative priority in number assignment, if it exists.
Evens:
x0- 20
x4- 15
x6- 10
x8- 6
x2- 4
Odds:
x5- 28
x9- 13
x7-10
x1- 6
x3- 6