The "hockey stick" graph refers to the results of Dr. Michael Mann (see "NIPCC vs. IPCC," Fig. 20, pg. 22).  In 1998, he presented a summary of his analysis of "proxy" data of the past 1,000 years, mostly tree-ring temperatures.  He showed a steady decline of temperature of the Northern hemisphere, without the "cooling" of the Little Ice Age (LIA), followed by a sharp temperature rise (the "blade" of the hockey-stick) in the 20th century.  The implication of his graph is obvious, but it is wrong in claiming that the copious emission of CO2 in the 20th century produced record warming in the decades before 1942.  Of course, the emission of CO2 grew after 1945, at the end of WWII, but the climate seemed to cool until about 1975.  (Apparently, the climate did not respond to CO2, as expected from greenhouse theory.)

However, Mann halted the use of his proxy data at the end of the 1970s decade and cleverly added the instrumental data for the period 1978-1998 from surface thermometers, which showed a sharp rise seen by surface thermometers that proved to be a complete illusion (some went so far as to use the word "fake").  He managed to suppress his own contemporaneous proxy data for that interval and has never revealed them.  To do so would have removed the sharp rise ("the blade") and ruined his "hockey stick" graph.

In other words, the "hockey stick" is a fictitious construct that has been called "Mike's nature trick" by some and has been "whitewashed" by the usual panels that were unaware of the background and implications.

The IPCC-Assessment Report-3 (2001) featured the "hockey stick" as the consequence as proof of human activity in the 20th century. Mann has since received many honors and lucrative government grants.       

This, despite the brilliant and persistent efforts and publications by profs Ross McKitrick and Stephen McIntyre.  Starting in 2003, these two Canadian statisticians were the first to show the "shoddy" analysis by Mann of his own data.  They even demonstrated convincingly that an input of pure noise fed into the Mann algorithm would lead to a hockey stick (REF).  Gradually, their scholarly critiques convinced others.  But there are still many "warmistas" who believe that the 20th century was the warmest in the last 1,000 years – despite strong observational evidence (REF) that the Medieval Warm Period (MWP) was warmer than the present,  and despite the existence of the well known LIA (Little Ice Age) of about 1400-1850 A.D.

In fact, the proxy data for the second half of the 20th century are now available.  They don't show Mann's sharp rise (REF; see also the tree-ring data of HTCS, Fig. 16, pg. 46 and the ice core data of NIPCC vs. IPCC, Fig. 24, pg. 24).  Up until now, Mann has refused to reveal his own proxy data, which include the crucial interval 1978-1997.  We are still waiting...  

S. Fred Singer is professor emeritus at the University of Virginia, founder and now chairman emeritus of the Science & Environmental Policy Project (SEPP).  His specialties are atmospheric and space physics.  He served as founding director of the U.S. Weather-Satellite Service, now in NOAA.  He is an elected fellow of major scientific societies.  In 2007, he founded the NIPCC (Non-governmental International Panel on Climate Change) (NIPCCreport .org).  For recent writings, see www.americanthinker.com/s_fred_singer and also Google Scholar.