Abstracts

Leon Battista Alberti and the Homo-

G.J. Hughes and D. J. Wright The EPOS (Eppendorf Patient Oriented System) selective chemistry analyser was evaluated using three chemistries: calcium, urea and gamma-glutamyl transpeptidase (GGT).
The dispensing system, photometer unit and temperature control system were all f6und to be excellent and complied with the manufacturers stated specifications. The betweenday and within-batch precisions were acceptable for all three chemistries as was the sample to sample carry-over. The linearity of the urea and GGT methods was shown to extend well above the reference ranges for each and the calcium method was linear between 1"5 mmol/1 and 3"3 mmol/1. The urea and GGT methods correlated very well with comparative methods when they were used to measure patient samples and they showed reasonable accuracy when used to measure reference sera although both had a small positive bias. The calcium method, however, showed a very poor correlation against a comparative method and the measurement of reference sera showed a positive bias at high values and a negative bias at low values which may have been related to a calibration problem.
Overall, the instrument was found to be well designed and simple to use. Chemistry's guidelines. The system's practicability has been studied, and the analytical units in terms of routine working, covering inaccuracy, imprecision, linearity and photometric drift have been evaluated. The sampler and the reagent dispensing system have been checked for inaccuracy and temperature control.
A rise of photometric imprecision has been observed with decreasing absorbance; linearity is good, drift is negligible and the imprecision of the pipette delivery system is acceptable.

Imprecision
(within-run and between-run), carry-over and relative inaccuracy have been studied under routine working conditions for several procedures: creatinine, total protein, ASAT, glucose and alkaline phosphatase. Carry-over was not detected, imprecision was low for enzyme activity measurement and acceptable for the rest of the constituents.
Relative inaccuracy was acceptable except for total protein, when compared with the measurements on the Ultrolab-Aurora. The protein problem could be caused by a calibration. Page 87 The use of queueing theory for planning automated analytical systems T. L. Pap and L. Leisztner Queueing theory is suitable for planning laboratories which need to analyse large series ofsamples. When the distribution of service time and inter-arrival time are not independent, then only a numerical solution to the problem is possible. In this paper, the authors describe an algorithm, which provides an appropriately simplified model of the laboratory. This algorithm is very useful for calculation of the queueing parameters. The function of the analytical system was modelled and the statistical data were processed by the help of computer program written on the basis of the algorithm. The development ofan immune complex by reaction between an antigen and antibody can be monitored by nephelometry or turbidimetry. The technique is used for the quantitation of specific proteins. The sensitivity of a method can be enhanced by using an antibody bound to a light-scattering particle. This approach can only be used directly for the measurement of polyvalent antigens.
Hapten molecules can be coupled to a core molecule such as a particle or a large protein molecule; this effectively produces a polyvalent antigen characterized by the hapten. If this reagent reacts with specific antibody to the hapten an immune complex is formed; formation of the complex may be inhibited in introducing a sample containing the hapten.