The first unambiguous detection of meteoroids impacting the night side of t
he Moon was obtained during the 1999 Leonid storm. Up to eight optical flas
hes were recorded with CCD video cameras attached to small telescopes on No
vember 18, 1999. Six impacts were videotaped by at least two independent ob
servers at the same times and lunar locations, which is perhaps the stronge
st evidence for their collisional nature. The flashes were clearly above th
e noise and lasted for less than 0.02 s. Although previous observational ef
forts did not succeed in detecting impact flashes, additional candidates ha
ve been reported in the literature. The evidence accumulated so far implies
that small telescopes equipped with high speed cameras can be used as a ne
w tool for studying meteoroid streams, sporadic meteoroids, and hyperveloci
ty collisions. In this review we discuss the various intervening parameters
for detectability of flashes on the night side of the Moon (geometrical ef
fects, contamination by scattered light from the day side, and properties o
f the meteoroids such as speed and flux of particles). Particular emphasis
is placed on the analysis of the observations in order to derive relevant p
hysical parameters such as luminous efficiencies, impactor masses, and crat
er sizes. Some of these parameters are of interest for constraining theoret
ical impact models. From a simple analysis, it is possible to derive the ma
ss distribution of the impactors in the kg range. A more elaborate analysis
of the data permits an estimate of the fraction of kinetic energy converte
d to radiation (luminous efficiency) if the meteoroid flux on the Moon is k
nown. Applied to the 1999 lunar Leonids, these methods yield a mass index o
f 1.6 +/- 0.1 and luminous efficiencies of 2 x 10(-3) with an uncertainty o
f about one order of magnitude. Predictions of visibility of the major annu
al meteor showers are given for the next few years. These include the forth
coming 2001 Leonid return, for which we estimate detection rates in the vis
ible.