1 Introduction

The Infrared Astronomical Satellite (IRAS) Point Source Catalog (PSC) contains a large number of Sources without counterparts in other astronomical catalogues. About 20$\%$ of these have $F{_\nu}(12~{\mu}{\rm m}) \geq F{_\nu}(25~{\mu}{\rm m})$ and lie within 7$\hbox{$^\circ$}$ of the southern galactic plane. Sources of this colour characteristic are expected to be mass-losing, giant stars of the M-S-C sequence having photospheric colour temperatures in the range 1500 - 3000 K. Although considerable efforts have been made to identify and observe selected groups of these in other wavelength bands since the publication of the PSC in 1988, a large number of these still remain unidentified/unassociated. Many articles have appeared on the IRAS colours of particular groups of IRAS sources, and position in colour-colour diagrams has been used to classify them as either C- or O-rich stars since they preferentially populate certain "occupation zones'' of the diagrams containing about 70$\%$ of the stars in each group (see e.g. Walker & Cohen 1988 and references therein).

van der Veen & Habing (1988, hereafter VH) studied the IRAS sources with circumstellar envelopes (CSE) and placed them into ten evolutionary groups based on location in the [25]-[60] vs. [12]-[25] colour-colour diagram. In their notation, the index [12]-[25] is defined as $2.5\times[ \log(F{_\nu}(25)/F{_\nu}(12))]$ and similarly for [25]-[60] where the $F{_\nu}(\lambda)'s$ are the IRAS flux densities at 12, 25, and 60 $\rm {\mu}{\rm m}$ uncorrected for colour dependence. Except when we discuss the [25]-[60] vs. [12]-[25] colour-colour diagram presented in Fig. 13, throughout the rest of the text, tables and figures (unless otherwise specifically stated) the IRAS magnitudes referred to by us are defined as $[\lambda]=-2.5\times[{\rm log}(F{_\nu}(\lambda)/F{_\nu} {_0}(\lambda))]$ where $[\lambda]$ is the IRAS magnitude of the source at the wavelength ${\lambda}$ in ${\mu}{\rm m}$, $F{_\nu}(\lambda)$, is the flux density in Jy at wavelength ${\lambda}$, and $F{_\nu}{_0}(\lambda)$ its zero magnitude flux density in Jy at wavelength ${\lambda}$ as given in the IRAS PSC Explanatory Supplement (IRAS PSC, 1988). Stars with O-rich envelopes form a sequence in the [25]-[60] vs. [12]-[25] colour-colour diagram. This has been interpreted as 1) due to evolution of mass loss rate (Olnon et al. 1984; Bedijn 1987; van der Veen & Habing 1988); 2) a sequence of increasing initial mass (Epchtein et al. 1990); and 3) due to the combined effects of increasing mass-loss rate and increasing initial stellar mass (Likkel 1990). VH showed that the evolutionary track of most of these stars in the [25]-[60] vs. [12]-[25] colour-colour diagram is a single-valued function of [12]-[25] colour. This track passes through regions I, II, IIIa, IIIb and IV of VH; see Fig. 13. There are, however, other stars with CSE that populate a much wider area of this colour-colour diagram. Some are those that have cooler envelopes due to higher emissivity of carbon dust in their envelopes. Thus, it has been claimed that the VH diagram serves as a useful tool for obtaining preliminary information on the evolutionary status of stars with CSE.

A major fraction of the unidentified sources of this study do not have best-quality (IRAS quality-3) flux density data at wavelengths longward of 12 $\rm {\mu}$m. Sources with $F{_\nu}(12~{\mu}{\rm m}) \geq F{_\nu}(25~{\mu}{\rm m})$are likely to emit fairly strongly in the optical-infrared region even if faint (as is expected of most of the unidentified sources). As spectral type data on these sources provide valuable information for an evolutionary classification of these sources, DJM carried out spectral classification of these sources using objective-prism plates collected by him. These plates were limited to the galactic latitude belt of $\pm$7$\hbox{$^\circ$}$. We present here results on approximately 10000 IRAS unidentified sources in the R.A. zone $5^{\rm h}17^{\rm m}$ to $12^{\rm h}06^{\rm m}$.