# A Modern Dwarf Stellar Effective Temperature Scale Based on 17,842
# Teff & MK Spectral Type Pairs 
#
# Eric Mamajek (U. Rochester) 

As of 20 January 2014, I have removed the original data table from
2010 from this page.

Short story: 

The 2010 table is obsolete. I encourage you to adopt the dwarf
spectral type vs. effective temperature (Teff) scale presented in
Table 5 of Pecaut & Mamajek (2013; ApJS, 208, 9; Table 5):
http://adsabs.harvard.edu/abs/2013ApJS..208....9P
http://cdsarc.u-strasbg.fr/viz-bin/Cat?J/ApJS/208/9
(and please cite that paper as the source). 

Long story:

A more refined analysis (taking special account of the 
intrinsic colors and measured effective temperatures of 
spectral standard stars, and vetting of those spectral standard
stars) has produced the dwarf effective temperature table
in Pecaut & Mamajek (2013; ApJS, 208, 9; Table 5):
http://adsabs.harvard.edu/abs/2013ApJS..208....9P
http://cdsarc.u-strasbg.fr/viz-bin/Cat?J/ApJS/208/9

The original analysis that went into the 2010 version of the Teff
table at this URL suffered from a few problems. Some of the problems
were: 1) it relied heavily upon a small number of spectral type
catalogs, some of which have systematic differences in spectral types
to more modern determiations (more on this later), 2) it was heavily
weighted towards stars that had more Teff determinations, many of
which were not necessarily good representatives of their spectral type
(e.g. metal poor or metal rich, cooler or hotter than other stars
typical for their type, etc.), 3) for some spectral classes
(especially hot stars), it was heavily weighted towards obsolete
catalogs that had many Teff determinations. For these reasons, a
different approach was taken. The goal was produce a dwarf Teff scale
appropriate for stars of roughly solar composition, where the Teff was
representative for the best spectral standards of those spectral
types. It was decided to attempt to evaluate this on a spectral type
by spectral type (and standard star by standard star basis).

The raw notes from that exercise can be found in the files in the
directory: http://www.pas.rochester.edu/~emamajek/spt/ We evaluated
the pedigree of the dwarf spectral standards that had been published
in the literature (from e.g. Johnson & Morgan, papers by Morgan,
Keenan, Gray, Garrison, Kirkpatrick, Henry, and others). Some older
standard stars had been largely ignored/abandoned after Johnson &
Morgan (1953), or reclassified by the original classifiers in later
papers, and newer standards have been introduced which largely
replaced the older standards (e.g. the Kirkpatrick and Henry M dwarf
standards). For example in the file "A0V.txt":
http://www.pas.rochester.edu/~emamajek/spt/A0V.txt I list all the
stars that have been previously mentioned as A0 dwarf standards, list
all their published spectral classifications, compile their colors and
published effective temperatures, grade the quality of the standards
(primary, secondary, [occasionally] tertiary, deprecated), and use all
of this to estimate the mean properties of A0V stars. This was a time
consuming process, but I'm confident that the mean values for each
spectral type that we published in Pecaut & Mamajek (2013) are about
as good as one can adopt given the limitations of the MK spectral
classification system. The MK spectral classification system has
changed very little over time, but it has migrated systematically and
crystallized a bit in a few key regions (especially among the A-type
stars, the F/G boundary, and late-K and M dwarfs).

For example, one may note that in the 2010 version of this Teff scale,
the mean Teff for G2V stars was listed as 5880K. This is 100K hotter
than the Sun, and the Sun is essentially "the anchor standard" for G2V
stars. The reason is that most of the G dwarf classifications that I
used in my original analysis rely heavily upon the Michigan Spectral
Catalogs (by Houk). The Houk survey began in the late 60s, and there
are some subtle differences (at the +-1-2 subtype level) among the F/G
standards that she used compared to the modern grid of F/G standards
(especially those of Keenan). For better or worse, the modern spectral
classifications (e.g. see especially large surveys by Gray, Corbally,
and collaborators) rely heavily upon Keenan's spectral standards -
which differed from previous classifications by Morgan, and Keenan
himself! Some particularly egregious examples of spectral standards
which "changed" over time are Eta Cas A (F9V by Keenan in post-1988
publications and most modern classifiers, but G0V according to Morgan
and Keenan before 1988!).  To illustrate this point a bit further:
Johnson & Morgan (1953; source of the "MK" standards) listed 3 G2V
standards: Sun, 16 Cyg A, and HR 483 (HD 10307). Keenan later
considered 16 Cyg A to be G1V, and HR 483 to be G1.5V. So the only one
of the JohnsonMorgan53 G2V standards to "survive" to Keenan's system
of standards was the Sun. Regardless, most "modern" spectral
classification of dwarf stars relies on standards from Walborn &
Morgan (OB stars), Morgan, Gray and Garrison (A/F stars), Keenan (G/K
stars), Kirkpatrick and Henry (M stars). Using classifications from
the 1950s-1970s for A/G/M stars for comparison with modern A/G/M stars
is probably unwarranted as the adopted standards stars have changed
somewhat -- one may introduce unwanted systematic effects at the +-1-2
subtype level.  As there have been some shifts in the standards
themselves, there have been some shifts in the mean properties as a
function of spectral type. The dwarf star Teff/color/BC tables in
Pecaut & Mamajek (2013) attempt to represent the mean properities of
dwarf stars of ~solarish composition on the "modern" system of
spectral standard stars.

Eric Mamajek
20 January 2014