5. Use of narrow band data (NB)

5.1. Temperatures for luminosity class III stars

To assess the reliability of the ISO calibration temperatures, we compare the published temperatures (Blackwell et al. 1990; Blackwell & Lynas-Gray 1994) using narrow band data, with the temperatures found using TCS narrow band data , discussed in the present paper. The mean difference between the published and the completely independent temperatures presented here is 0.05 per cent with a standard deviation of 0.75 per cent. This is reasonably satisfactory, and we therefore consider it justifiable to take the average of the three (or occasionally two) temperature determinations for Class III stars, i.e. those published and the new NB temperatures, and relate them to photometric colours. The results are presented in Table 4 (click here), which shows polynomial coefficients together with respective percentage standard deviations from the regression curve for (B2-V1), (b-y), (V-K) and (B-V), all corrected for interstellar extinction. The table suggests that (B2-V1) may be considered a good indicator of temperature, as Hauck (1994) has emphasised, and that it is clearly useful for obtaining reliable photometric estimates of temperature. Figure 2 (click here) shows the close relation between temperature and .

  
Figure 2: Plot of temperature of Class III stars against where K is on the Johnson scale

5.2. Temperatures for luminosity class V stars using narrow band data

A comparison between the average of published temperatures (Blackwell et al. 1990; Blackwell & Lynas-Gray 1994) with the new ISO calibration star temperatures shows a mean difference of 0.43 per cent with a standard deviation of 1.06 per cent. This suggests that the two batches of data may be combined. The regression constants for the combined data are shown in Table 4 (click here) together with standard deviation values. The standard deviations are comparable with those for Class III stars presented in the same table. Figure 3 (click here) shows the good relation that exists between temperature and . There is no significant improvement in standard deviation for either class when the chosen temperature range is smaller.

  
Figure 3: Plot of temperature of Class V stars against where K is on the Johnson scale

A comparison of polynomials for Class III and V star temperatures obtained using narrow band data shows a real difference between the derived temperatures for a given colour, of order 5 per cent.

   

Class M p A B C D

III 1 (B2-V1) 7626 -6841 5514 -2293 0.82

(b-y) 8757 -11242 10676 -4616 1.04

(V-K) 8970 -2970 626 -53 0.73

(B-V) 8925 -7664 4867 -1339 1.55

V 1 (B2-V1) 7675 -6847 7819 -7967 0.67

(b-y) 8870 12789 20418 -19250 0.67

(V-K) 9137 -3496 1063 -158 0.82

(B-V) 8973 -7908 6087 -2575 1.04

III 2 (B2-V1) 7617 -5847 2881 -430 1.51

(b-y) 8512 -8871 4743 -343 1.44

(V-K) 9052 -3032 672 -66 0.58

(B-V) 8546 -5616 2139 -234 1.15

V 3 (B2-V1) 7738 -7832 7594 -4414 1.04

(b-y) 9017 -12771 14831 -9791 1.27

(V-K) 9352 -3583 974 -117 0.57

(B-V) 9139 -8853 6558 2163 1.54

V 4 (B2-V1) 7131 -4273 675 2.30

(b-y) 8151 -7025 1815 2.31

(V-K) 8831 -2724 501 -33 0.73

(B-V) 7404 -3029 286 2.74

III 5 (B2-V1) 7627 -6369 4229 -1325 1.14

(b-y) 8579 -9288 5005 0.96

(V-K) 8723 -2516 352 1.30

(B-V) 8826 -7197 4400 -1201 0.70

V 6 (B2-V1) 7876 -7806 4886 2.23

(b-y) 9054 -11539 7535 2.59

(V-K) 8326 -2153 260 0.78

(B-V) 9133 -7753 3468 1.54 Notes: 1 - Narrow Band Average 2 - Average of TCS and Narrow Band

Table 4: Regression Coefficients A, B, C, D for relations between stellar temperature T and values of designated photometric parameters p: T=A+Bp+Cp²+Dp³; is the corresponding percentage standard deviation in temperature and M the measurement code noted below