In this case, since the measurement result is the addition of all nine measured values, it is correct to add the bias uncertainty of each measurement for each resistor in the series of nine; as the bias uncertainty of the Ohmmeter is correlated at +1 (a 'plus' error indicating meter times 9 measurements = 9 plus errors added together). You should also include the calibration uncertainty of the instrument in the correlation method for each measurement of the series of resistors (Cal Uncertainty times 9). Then RSS with the independent uncertainty components with the added components, as mentioned. You are correct, except for environment, it too has a correlated effect, if it causes a plus error due to temperature, that error is persistent in all nine measurements.
James D Jenkins