WizardOfBoz

So I was thinking about this, initially disagreeing, but then... Normally you tighten a bolt down against a hard surface, and the purpose of the torque reading is to stretch the bolt to an optimal degree. So when you tighten a connecting rod to a crankshaft, the clamp seats on the connecting rod and torquing stretches it. Because there is variable friction and people use different preps (like locktite, anti-seize, grease) torquing is not the most reliable way to get a precise stretch. Using the connecting rod as an example again, really precise work is often done by measuring the untightened bolt length with a micrometer, and then making sure that the bolt is stretched to a precise, optimal value. A bolt that is less tight actually provides a weaker joint than a bolt that is slightly over-torqued (see Carrol Smith's book on Fasteners). But for tightening two parts together tightly, the torque value is determined by the bolt material, thread pitch, and what has prep stuff been applied.

This all changes if the bolts tighten the clamp onto the bar and the two halves of the clamp don't touch. Then the whole point of the torquing exercise is to get a precise clamping force on the bar. An ideal design would use adjust the bolt size and number so that the torque values were identical - the bolt joint will be stronger with the ideal amount of stretch.

Going back to the OPs question, though, I would use the torque values on the part. They are probably set up for optimal strength. If I had replaced (for example) steel bolts with Ti, I'd check to ensure that the torque values were within the acceptable range of the Ti bolts.

I like Calvin Jone's (Park Tool) videos. Here's one you may like, too: