Turbulence is ubiquitous in astrophysics, ranging from cosmology, inte
rstellar medium to stars, supernovae, accretion disks, etc. Large scal
es and small viscosities combine to form large Reynolds numbers. Becau
se it is not possible in a single article to review all the above scen
arios, we limit ourselves to stars, in which thermal instabilities giv
e rise to turbulent convection as the dominant heat transport mechanis
m. (Accretion disks, where shear instabilities dominate the outward tr
ansport of angular momentum, will be the subject of a second article,
planned for Volume 31.) Because of the lack of a satisfactory theory,
turbulence constitutes a bottleneck that prevents astrophysical models
from being fully predictive. Because continued use of phenomenologica
l turbulence expressions would make astrophysical models perennially u
npredictive, a way must be found to make astrophysical models as progn
ostic as possible. In addition to the difficulties brought about by tu
rbulence, astrophysical settings introduce ''malicious conditions,'' o
f which the most refractory to a satisfactory quantification are compr
essibility (caused by the large density excursions that characterize c
onvective zones in stars) and rotation. Basic understanding of how the
y affect turbulence in general is still rather sketchy. Reasons for th
e choice of stars and accretion disks as prototype examples are the fo
llowing: The underlying instabilities are very basic; laboratory and d
irect numerical simulations data help constrain theoretical models; an
d new observational data, especially from helioseismology, help discri
minate among different models with unprecedented accuracy.