Step 1: Recall the accepted installation rules for a measuring weir.
For a weir to give an accurate head-discharge relationship, the approach channel must be uniform and free of turbulence, the crest must be raised well clear of the channel bed so the approach velocity is negligible, the head must be measured far enough upstream that the drawdown curve near the weir does not distort the reading, and the nappe, the sheet of water falling over the crest, must fall freely with air able to reach underneath it.
Step 2: Test statements A and B against these rules.
Statement A says the velocity of approach should not be considered, but a real approach velocity adds kinetic energy to the flow and is either minimised by raising the crest or explicitly corrected for, never simply ignored, so A is false. Statement B says the crest should be placed closer than 2H to the channel bed, but the actual rule is the opposite, the crest must be at least about twice the maximum head above the bed so the approaching flow has settled, so B is also false.
Step 3: Test statements C and D.
Statement C says the head should be measured upstream at a distance of at least four times the head from the weir, matching standard practice of gauging the head where the surface is still level and unaffected by the drawdown near the crest, so C is true. Statement D says the falling water should not have free flow, but a ventilated, freely falling nappe is a strict requirement for the standard weir formulas to hold, so D is false.
Step 4: Combine the results.
Only statement C survives as an accurate installation requirement.
\[ \boxed{(C) only.} \]