M. Járó - L. Költő szerk.: Archaeometrical research in Hungary (Budapest, 1988)

Dating - BENKŐ Lázár: Thermoluminescence dating of Hungarian archaeological sites (potteries, hearths, calcite)

Capsules were buried for 1 to 2 years in the undisturbed soil of the excavations at Tiszapolgár—Basatanya. The mean capsule dose-rate was 0.97 mGy/y hence the true value is 1.05 mGy/y. Beta dose-rate (Dß). The saturation water content, as measured before cmshing, is W = 0.10. The plastic tube assembly was previously calibrated by diluted radioactive ores of specified uranium and thorium content and by using K

2

C0

3 as well. On the basis of the radioisotope concentrations, the ratio between the phosphor dose and the point dose within the radioactive matrix was found to be 0 51. The beta attenuation for the quartz grains of about 0.1 mm should be taken into account. This is a factor by which the average dose to the grain is smaller than the dose to a fine-grain (point dose). A factor of b - 0.90 is recommended. According to the above statements the actual beta dose-rate for sample No.5335.116. is 2.96 mGy/y and therefore the total (gamma + beta) annual dose is 4.01 mGy

. TL age (A). The TL age can be calculated from the following equation: P 23.2 X 10

3 mGy ,_

0

. _ = . J = 5786 years. D^Dß 4.01 mGy Assessment of error limits. The procedure is to calculate the percentage error in the date corresponding to a given error in each of the quantities on which the date is based and then to obtain the overall error as the square root of the sum of the squares of the individual errors. Random errors arise partly from the measurement uncertainties in Q (6Q = 0.68 Gy), partly from uncertainties in beta and gamma TL dosimetry measurements (in our practice SDß = 5D7 = ±4%). Hence the corcesponding percentage errors in age are given by 2 / 100 ő Q \

2 . /100 5 I and respectively, where fß = 0.74 and fx. = 0.26 are the fractional components of the dose-rate. The random error can be written as a

r = (8.74 + 9.86)

1/2

. Systematic errors are mostly due to calibration uncertamties (ai = 27.4), parameter uncertainties (ol = 15.4), radon escape (o\ = 4) and wetness (a

6 = 13.18). Hence a

s =(59.98)

1/2 and o = {o\±ol)

in =8.86%. The quantity e = a A/100 gives the standard error. In the present case e = 513 years. When several samples from a context are investigated, the best value for the age of the context is obtained by weighting the individual ages. This has been done for the site