I'm not into wood, but I would suspect the lumber is Machine Stress Graded (MSG). Stronger members are placed in the outer fibres where the flexural stress is greater.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

The laminations are each just pieces of structural lumber, and they'll be graded in a similar fashion (though as dik mentioned, many of them are machine graded rather than the visual grade lumber that is commonly available to consumers). So you have allowable axial stress in each lamination and allowable shear stress in the glue at the joints. They can select lumber of various grades to match the flexural stress profile, even to the point of having LVL laminations at the extreme edges where stresses are higher and cheap 'filler' wood near the neutral axis. Then you have balanced and unbalanced layups - if it's a floor beam that will only ever have positive bending, then you can save a lot of money in a large building by placing laminations with high tension capacities at the bottom of the layup and selecting the top ones for compression only (that's unbalanced). For continuous beams or roof beams where you may have negative bending, a balanced layup gives you tension capacity on the top as well. It's also good to do balanced if you're doing a low-end job where you can't be sure if the laborers will read the "THIS SIDE UP" stenciled on the beam or if it's going to be a stock order - I believe most yards stock balanced since they are more versatile but can buy cheaper unbalanced layups on a large order.

Pham is correct on nearly all accounts, save the last statement. If your project is on the west coast, nearly all lumber yards stock unbalanced layups, however, balanced, or on the west coast, V8 glulam, is nearly always special order, with associated lead times. This may differ on the east coast, where I believe Pham is located.

5th percentile isn't exactly correct, but the design values are based on an adjusted 5th percentile, in essence, per the ASTM standard, you divide the 5th percentile by, if I remember correctly, either 2.1 or 2.37, depending on the design value being evaluated. Part of that is an adjustment for load duration (generally 1.6 due to length of sample test time), and the other an adjustment to reduce nominal strength below the 5th percentile to ensure sufficient capacity of the design member.

ChorasDen - thank you for the correction. Yes, I am on the east coast - in Virginia.

Thank you all, very interesting and all of them good points - trying to learn as much about wood as I can. I was able to get a hold of ASTM D3737 and spent the day looking at the procedure. Seems like, in general, the allowable properties for glulam are based on stress index values (based on ASTM D2555 as described in my original post), which are then multiplied by stress modification factors, and then modified for specific conditions.