Let's began in the way how the isocrones are computed. The basic idea was quoted in:

The Astrophysical Journal, 203: 52-62, 1976 January 1

Evolutionary Synthesis of Stellar Population in Elliptical Galaxies I. Ingredients, Broad-Band Colors, and Infrared Features

Beatrice M. Tinsley
James E. Gunn

In Section II. Stellar Ingredients, subsection b) Evolving Dwarfs and Subgiants

In this paper it is shown the first problem to obtain isochrones: the definition of equivalent evolutionary points and the use of homology relations. Some usefull references for understand what homology means are:

• Kippenhahn & Weigert 1990 STELLAR STRUCTURE AND EVOLUTION (Springer-Verlag).
• Mowlavi et al. (1998) A&A 335, 573 (Appendix)

In sumary, an homology transformation gives us a scale relation of a property x of a star (L, T_eff, R, _ms) with their Mass, M, mean molecular weight, , nuclear energy production, , opacity coefficient, , and Hydrogen abundance in the core, X, for the same configuration. As an example, for the CNO burning in the Main sequence, Kramers opacity law and radiatie transport configuration the relations are:

Other configurations means, mainly, a changue in the exponents. These relations, together with , allow us to define equivalent evolutionary points: points where , (structure) and , and X (burning state) have the same value. So, for these reasons,

1. Tracks are tabulated following the abundance of the central burning element (H, He, C)
2. Interpolations must be done in the log M - log x plane.

(Un)fortunatelly, this simple squeme is not valid for all the mass range because the internal structure of the star changues. Additionally, the homology relations are a good aproximation only for Main sequence stars. Then, for these reasons there are tracks with different initial masses and not only an individual track.

However, it is assumed to be a good aproximation for all the evolutionary stages (it is the most simple assumption and it is in fact implicit in all the synthesis codes, as far as I know). It woul be valid if

• The stellar tracks are close enough in the HR diagram.
• There is no discontinuities in the stellar evolution.

i.e. when adyacent tracks ``resemble'' each other

However. since the evolution of Post Main Sequence Stars (PMS) are quite important in the integrated ligth of an stellar cluster, there are additional ways to compute the PMS evolution.

1. Obtain directly the isochrone from interpolations of the tracks (as I explained before). This is the most usual in models for star forming regions:
   • Adventages: It obtain easily the properties of individual stars; i.e. for each star in the isochrone, their physical properties are known and the integration with the IMF are quite easy.
   • Disadventages: It does not asure that the mathematical interpolations reproduce a real physical situation: i.e. some of the stars in the isochrone may not be physicaly consitent. As example:
     1. There are no homology relations for compute mass lost rates.
     2. There are no hology relations for compute the surface abundances (needed for obtain the WR population and for the ejection rates of massive stars for chemical evolution models).
     3. Internal structure (and lifetimes) changue with mass lost (i.e. homology relations fails).
   
   
2. Use the Post-Main Sequence part of the Isocrhone from stellar tracks of individual stars:
   • Adventages: The isochrone is physically self-consistent.
   • Disadventages: It is needed aditional assumptions to populate the isochrone, like the Fuel Comsumption Theorem (Buzzoni 1989, ApJS 71, 817).
   
   Theoretically both methods should be equivalent, as the next Figure shows. Even more, it can be demostrated that there are a critical mass point in the isochrone where the following points in the isochrone coincides with the track of an individual star.

The problem with isochrones obtained from twice mass lost rates tracks are:

1. A technical problem in the interpolations, since the tracks (especially the WR phase) is quite dependent on the stellar mass. Even more, the lifetime of the stars is not a monotonuous function of the initial mass (stars with 120 Mo "die" latter than stars with 60 Mo!).

2. Observational mass lost rates are even lower that the one from de Jager, Nieuwenhuijzen & van der Hucht (1988 A&AS 72, 259) assumed in the tracks with standrad mas lost rates, has has been showed in Kudritzki & Puls (2000 ARA&A 38, 613) (see also here)

3. Evolutionary tracks with twice mass-lost rates do not agree with observations of real stars (Ph. Massey in the proceeding of IAU Symp. 212)

4. The only argument for the use of tracks with Evolutionary tracks with twice mass-lost rates is that by ussing evolutionary synthesis models, that tracks looks to fit the WR ratios in high-metallicity systems, whereas in low metallicity systems it is always needed the presence of binary systems (Maeder & Meynet 1994 A&A 287, 803), but if the isocrone computation (and hence the results of evolutinary symthesis models) is not correct, there are no place for the use of such tracks.

In sumary:

The use of tracks with twice mass lost rates is unphysical

Finally, the isochrone computation is also dependent on how the time interpolation is done. Problems that arise related with the time interpolation has been shonw in Cerviño, Gómez-Flechoso et al. (2001 A&A 376, 422), as example, the computation of the SNrate and the kinetic energy emitted by the starburts. I refer to that reference for more details.

The final point that must be addressed is the impact of the use of different atmosphere models in the synthesis codes. Such a subject is also quite extensive, and I am not going to address it here (mainly, because I have no time to address it properly at this moment)... sorry.