Abbildungen der Seite



It had long been assumed that asteroids were gray reflectors of the solar spectrum and they have been used from time to time as comparison stars. Bobrovnikoff (1929) first questioned this premise and attempted to measure the characteristics of asteroid spectra. He compared microphotometric tracings of photographic spectra with G-type stars; he concluded that (1) he was observing reflection spectra with no emission features, (2) that Ceres and Vesta lacked any major absorptions in the visible like those of Jupiter, (3) that asteroids have relatively low reflectivity in the violet and ultraviolet, and (4) that there were differences between asteroids. Bobrovnikoff’s tracings seem to show definitely that Pallas is relatively more reflective near 0.4 pm than other asteroids studied. But Watson (1938) regards many of Bobrovnikoff’s conclusions as uncertain because of a lack of standardization of the spectra. Certainly there are some discrepancies with recent photoelectric data for some asteroids discussed by Bobrovnikoff.

Microphotometric tracings of spectra of three asteroids by Johnson (1939) yielded the incorrect result that these asteroids were substantially bluer than the Sun. Recht (1934) reached a similar erroneous conclusion from a more extensive study of the color indices of 34 asteroids obtained from magnitude measurements on normal photographic and panchromatic plates. Recht's measurements have been criticized by several subsequent writers. They show a large scatter because, among other reasons, the measurements of the two colors were often made from plates taken on different nights, and there is a strong correlation between the color index derived by Recht and the apparent magnitude of the asteroid—such a correlation being indicative of a spurious systematic error in the photographic measurements. There is little if any agreement between Recht's color indices and recent UBV photoelectric photometry. Watson (1940) obtained more realistic color indices for seven asteroids, but their reliability is difficult to gage.

Perhaps the most ambitious and reliable of the early photographic colorimetry is that of Fischer (1941). Though Fischer's data show less scatter than Recht's, the random errors are nevertheless uncomfortably large. Of the 30 asteroids for which Fischer obtained color indices, a fair number have photoelectric B - V colors that correlate reasonably well in a relative sense with Fischer's values. In figure 1, Fischer's color indices have been rescaled and plotted so that their mean and range match the photoelectric values, but no absolute calibration is intended. It is probably true that most of Fischer’s bluer asteroids are in fact bluer than his redder ones, but finer distinctions probably have no meaning. Fischer reported statistically significant correlations between color index and two related orbital characteristics: semimajor axis and Jacobi constant. The correlation is in the same sense as evident in subsequent photoelectric work (see later section), but one should be aware of the potential

[ocr errors]

Figure 1.-Asteroid colors from photoelectric and photographic photometry. (A) B V, U B colors for asteroids summarized in Gehrels (1970). Triangles indicate those asteroids that have also been observed in the present authors’ spectral reflectivity program. The line is the stellar main sequence and its intersection with the arrow in the lower left is the UBV color of the Sun. (B) B V colors for asteroids for which no U color has been measured (Gehrels, 1970). (C) Approximate scaling of photoelectric color indices of Kitamura (1959). (D) Approximately rescaled photographic color indices of Fischer (1941); probably only the gross distinction between asteroids on the left and to the right is meaningful for Fischer's colors. Also see figure 1 of the paper by

[ocr errors]

in this early photographic work for systematic apparent-magnitude-dependent

errors that themselves would be weakly correlated with semimajor axis.

The largest sample of asteroids for which photographic color indices have been determined is contained in the work summarized by Sandakova (1962). At the time of writing, we have not been able to obtain the original published description of these data and their reduction. It is worth noting that seven asteroids with color indices reported in Sandakova (1962) show poor agreement with recent UBV work and that very large scatter is evident in the data for the 10 asteroids reported. Sandakova reports no correlation of colors

[ocr errors]

in the complete sample with a, but a large difference in color between asteroids with unusually small and unusually large orbital Jacobi constants.


An early photoelectric program to study asteroid colors was carried out by Kitamura (1959) in the mid-1950's. Forty-two asteroids were measured with a 1P21 photomultiplier in two colors with effective wavelengths somewhat longward of the standard B and V colors. From a graph presented by Kitamura of the color indices of six stars with known B - V colors, it is possible to make an approximate conversion of his color index to B - V. The resulting values have a slightly redder mean and greater range than B - V colors obtained by Gehrels, Kuiper, and their associates, so we have applied some corrections to Kitamura's colors for plotting in figure 1. The several cases of multiple measurements of the same asteroid show small scatter in Kitamura's data and the agreement for those asteroids for which B- V colors are known is good. Kitamura reports negative attempts to correlate his color indices with the proper orbital elements, magnitude B(1,0), and rotation period. Though his figures show no correlation with B(1,0) or mean motion, there appears to be a definite correlation with proper eccentricity e'. The sign of the correlation is such as to amplify the expected correlation of the Jacobi constant with respect to a correlation with a His table also shows a possible correlation of color index with extreme a, such that asteroids with a > 3 AU are bluer than those with a < 2.3 AU (but the statistics are poor).


Since the mid-1950's Gehrels, Kuiper, and their associates have published a series of papers on photoelectric photometry of asteroids in the standard UBV system. Gehrels has published a table summarizing these results (Gehrels, 1970) and we have plotted them in parts A and B of figure 1. The plotted colors include the small corrections made by Gehrels for Feddening with phase; he used lunarlike phase relations, the applicability of which to asteroids has been largely untested. The consistency of most of the UBV data is quite good, and most of the plotted asteroids are probably known to at least 0.05 mag in both colors. Of course, there are rarely sufficient data to determine the ranges of variation in color with rotation and phase for the individual asteroids, and such variations would contribute to the scatter. One asteroid, 1566 Icarus, has a typical B - V color but its U- B value is so large that the point is off the scale of the figure.

There is a fair spread of asteroid colors evident in the figure with a trend somewhat redder than the stellar main sequence. There is a major clumping around (B - V, U- B) = (0.83, 0.4) and a lesser one near (0.7, 0.25). There is some spread of the main clump both to the upper right and to the left. The numbers of several asteroids for which only B - V colors exist are plotted in part B of the figure. In sum, there is a general dearth of asteroids with B - V colors near 0.75. For purposes of comparison, Kitamura's rescaled colors are plotted in part C of the figure and Fischer's rescaled colors in part D. In general, these three sets of data show fair agreement, but there are discrepancies. Parts A and B of figure 1 are replotted in figure 2 showing the distribution in color of five groupings by asteroid semimajor axis. A correlation is evident, though it is due almost entirely to the extreme values of a. Ten of the 13 asteroids with a > 3.0 have B - V's 0.8 whereas none of the five asteroids with a < 2.3 is so blue. Asteroids with 2.75 × a < 3.0 show the greatest range of colors. If several times as many asteroids could be plotted, we might begin to see statistically significant clusterings of a values in the plane of figure 2, but it is premature to draw strong conclusions from the present sample.

[ocr errors][ocr errors][ocr errors][ocr errors]

Figure 2. —Color of asteroids from parts (A) and (B) of figure 1 plotted in five groups according to size of semimajor axis.

We have noted that the correlation of UBV asteroid colors is somewhat stronger with the Tisserand invariant against perturbation by Jupiter (Jacobi constant) than with a alone. Nevertheless, the correlation is far from perfect and there are several extreme exceptions.


McCord and his associates have been undertaking a program of measuring the spectral reflectivities of all major solar system objects from 0.3 to 1.1 p.m, and out to 2.5 pm when possible. After enticing results were obtained for Vesta, a program was begun to look at as many other asteroids as possible. This program constitutes the major portion of Chapman's doctoral dissertation, now in preparation. Although a program of strictly asteroid photometry has not yet been funded, telescope time has been available for asteroid observations during hours when other objects of high priority were below the horizon. To date we have observations of some sort of 32 asteroids, of which 12 have been partly reduced and will be discussed later.

McCord's (1968) double-beam photometer has been used in making observations of asteroids in a variety of modes on several telescopes at Mt. Wilson, Mt. Palomar, and Kitt Peak. A set of 24 narrowband interference filters from 0.3 to 1.1 pm are used concurrently, sometimes in a spinning-filter-wheel mode (3 rpm), and sometimes incrementally over a period of about 1 hr. The sky is observed in the second beam of the photometer with a 10 Hz chopping system and is subtracted from the signal. For some runs an S-1 phototube is used over the entire range 0.3 to 1.1 pm, whereas for others the S-20 is substituted for the wavelengths to which it is sensitive. Most of the data reported in this paper were taken with the S-1 tube alone. A pulse-counting data system is used. Air-mass corrections are determined from observations in each filter of the standard stars of Oke (1964) by taking values at equal air mass and correcting for time-dependent changes. The data are reduced to spectral reflectivity using the stellar standardizations and the solar spectrum of Labs and Neckel (1968). However, integration over solar spectral lines and bands with square-wave filter response produces error, especially near large solar lines in the ultraviolet. All standard stars are ultimately tied to o-Lyrae by Oke and Schild (1970) and, therefore, systematic errors in o-Lyrae's flux distribution will affect our results. However, theoretical models for o-Lyrae and observations presently agree to within a few percent over our spectral range. Deviations of a few percent of particular filters from the general trends that are observed for all solar system objects are smoothed out. All sources of error are very small, however, so the accuracy of our standardizations is a few percent, except for one or two ultraviolet filters. The relative comparisons between solar system objects are even more precise. The reflectivity curves are scaled to unity at 0.56 pm for purposes of comparison.

Reflectivity curves obtained in this manner bear some relation to UBV colors but provide much more information. Asteroids with identical UBV colors may differ greatly in the red and near infrared regions where important absorption bands are common. In fact, the details of spectral reflectivity curves in the 0.3 to 0.6 pm region can differ somewhat for asteroids with identical UBV colors, although the overall trends must correlate. Thus, far more

« ZurückWeiter »