Observed period-luminosity sequences

Infrared photometry in J and K was obtained on 7 December 1995 for all the giant stars in the half degree study area using the infrared camera CASPIR (McGregor et al. 1994) on the 2.3m telescope of the Australian National University. The K magnitudes from these observations were combined with the periods of the variables obtained from MACHO data (Alcock et al. 2000) and plotted in the (K,) plane in Fig. 1. Measurement errors in K are estimated to be less than 0.03 magnitudes. Since the K magnitudes are single-phase measurements, there will be a scatter introduced by variability of the stars. For the Mira variables, which are the largest amplitude long-period variables (LPVs), the total K amplitude is 0.6 mag. on average (Feast et al. 1982, Wood et al. 1983) so that the individual K measurements for these stars might be expected to deviate from the mean by up to 0.3 magnitudes. For non-Mira LPVs, which have smaller amplitudes (see below), the deviations from the mean magnitude should be much less.

Figure: The variable LMC red giant stars plotted in the (K,) plane. Five sequences labelled A...E can be seen (see text). The positions of the tip of the first giant branch (FGB) and the minimum luminosity for thermally pulsing AGB (TPAGB) stars with M 1 are indicated by arrows. The solid and dashed lines are the K- relations from Hughes & Wood (1990) and Feast et al. (1989), respectively. Solid circles correspond to stars with J-K > 1.4 and they are assumed to be carbon stars. Other stars are assumed to be oxygen-rich M or K stars.

The most striking feature of Fig. 1 is the group of four parallel period-luminosity (PL) sequences (A,B,C,D) brighter than 13 and the sequence E which consists of stars on the first giant branch (it also has an extension between sequences C and D - see Wood et al. 1999).

It has been known for many years that the Mira variables in the LMC lie on a single PL sequence in the (K,) plane (Feast et al. 1989; Hughes and Wood 1990). Furthermore, Wood and Sebo (1996) had found evidence for a second sequence at shorter periods (Bedding and Zijlstra 1998 have shown that the second sequence also appears to exist among the local semi-regular variables). The Mira sequences of Feast et al. (1989) and Hughes and Wood (1990) are shown in Fig. 1 and they clearly correspond with the sequence labelled C. The largest amplitude variables also lie on sequence C, providing more evidence that these stars are Mira variables. The second sequence of Wood and Sebo (1996) corresponds to the sequence labelled B in Fig. 1. These stars are of smaller amplitude than many of those on sequence C and would be known as semi-regular (SR) variables in our Galaxy. Note, however, that many of the stars on sequence C are also of relatively small amplitude (<1.5 mag. in V) and would also be classified locally as SR variables. The stars of sequence A are of very low amplitude (<0.2 mag.) and would probably be classified locally as SARVs (small amplitude red variables): these stars were not previously known to fall on a well-defined PL sequence. In summary, the new MACHO data clearly highlight the previously-known PL sequences and show several additional sequences.

© Copyright Astronomical Society of Australia 1997