For ensuring the traceability and uniformity of measurement results, the main objectives of national metrology in chemistry are to calibrate and verify measuring instruments, to evaluate the uncertainty of measurement results, to intercompare the analytical results etc. The concept of traceability has developed recently in chemical measurements, thus, an attempt to implement the principles of metrological traceability especially by appropriateness calibration using the composition certified reference materials (CRMs), are underlined. The paper presents some aspects and practices in the field of spectrometric measurements regarding the metrological quality of the traceability by calibrating the instruments using suitable and reliable CRMs. The uncertainty of the results, as a measure of the reliability that can be placed on them, has been adequately described in different documents, and, as a consequence, some examples of evaluating the measurement uncertainty are described. The relationship between uncertainty and traceability, as two fundamental concepts of metrology which are intimately linked, is underlined.

The thermal and mechanical stresses resulting from a short-circuit fault in an electric circuit strongly depend on the circuit power factor. For this reason, the IEC standard related to short-circuit test indicates, for each current range, the value of the power factor with a defined tolerance. A method for the measurement of the power factor based on the analysis of the digitally recorded short-circuit current is briefly described. The method is checked by applying it to the evaluation of the power factor measurement value and uncertainty in simulated test conditions. Attention is focused on the harmonic distortion of the supply voltage: the obtained results show that the method can be applied also in this case, but with a decrease of the measurement accuracy, which depends both on the amplitude of the distortion and the value of the power factor.

Rules of evaluation of uncertainty as a measure of measurement accuracy are presented in the well-known "Guide to the expression of uncertainty in measurement". There are a few examples of uncertainty calculations in literature. However, all these examples refer to the case when the output quantity is determined on the basis of known input quantities by an analytical formula. In this paper a case when the output quantity is determined on basis of known input quantities by solving numerically a set of equations is discussed. It was shown that the rules recommended in the Guide cannot be used directly to calculate the standard uncertainty of the output quantity in this case but the calculations can be done using Monte Carlo simulations. Next, a software package that enables quick determination of the combined standard uncertainty or the extended uncertainty not only in case of analytical models of measurement process but also when the measurement result is calculated by numerical methods or algorithms is presented.

Metrology, in general, and issues such as traceability and measurement uncertainty, in particular, are increasingly demanded when checking the safety of a wide range of food and agricultural products, in chemical manufacturing, environment protection, clinical chemistry as well for regulatory purposes. In many chemical quantitative analysis, the evaluation of measurement uncertainty requires the practician to look closely at all possible sources of uncertainty. Among them, data provided in calibration and correlation between the input quantities call for insight based on experience and general knowledge. Some aspects related to the estimation of measurement uncertainty and of calibration uncertainty in this field are presented.
Despite the greater attention given in the recent years to covariances and correlations in metrology, they still seem to be undervalued in chemical measurements based on relative or comparative methods of measurement. Consequently, the paper describes the applicability of the variance- and covariancepropagation law in the particular case of spectrometric analysis. An example showing the magnitude of the correlation between output estimates is discussed.

In the proposed paper a possibility of automation of calibration of liquidin- glass thermometers is discussed. Temperature readings are taken with the use of a measuring system that consists of a standard calibration facility, the video camera, and the personal computer that executes the program for image processing and analyses. Once the reading of temperature is extracted from an acquired image, this value can be passed to any other data processing, which includes the comparison with the values from the reference thermometer and determination of total uncertainty. Data achieved in this way can be directly used to prepare a calibration certificate.