[NOR SENT]

  "The other day I paid a visit to Dr. Randi Geller of Palo Alto,
  CA. I wanted to see for myself the remarkable Spinclone People
  Breeder that he described at the Biorythmics Conference 
  in Tahiti last Spring.
  
  "I was quite surprised by the size of the machine. This is not a lab
  benchtop device: it is a full-scale industrial installation that
  completely fills a largish building, about the size of a small
  supermarket.
  
  "They call it "People Breeder" for the lack of a better term, but in
  design and operation it does not resemble a breeder as much as an
  enclosed roller coaster or joy ride.  Try to imagine Disneyland's
  "Pirates of the Caribbean", redecorated according to the tastes of a
  chemical engineer.
  
  "The people to be cloned enter in a continuous stream through a door
  on the left side of the building, where they are helped by
  Dr. Geller's assistants into small single-person rail cars.  Each
  car then zips through the building at considerable speed, following
  some complicated three-dimensional path, fulls of twists, turns,
  bumps, knots, and jerks.  (More details can be found in Dr. Geller's
  patent and papers.)  Eventually, the cars stop by the exit door,
  where the riders are helped out of their seats and leave the
  building.
  
  "The cloning of people happens somewhere along that ride.  The
  physics of the phenomenon is still not completely understood, but
  Dr. Geller believes that it has something to do with the stretching
  of spacetime by the high speed of the cars, and possibly by the
  peculiar pattern of centrifugal and gravitational forces experienced
  by the riders.
  
  "Anyway, these theoretical questions are largely irrelevant.
  Dr. Geller is a practical man, and all that matters to him is the
  fact---confirmed by umpteen experiments and attested by many
  satisfied customers---that more people leave the building than
  go into it.
  
  "During my visit I witnessed three such experiments, two of which
  generated unmistakable amounts of excess people.  We are not talking
  about 1% discrepancies, which the skeptics can always attribute
  to stragglers or counting errors.  We are talking about 50% to 100%
  excess people, meaning hundreds of new people cloned per hour; and
  these levels can be sustained for years on end.  Needless to say,
  excesses of this magnitude cannot be attributed to measurement
  errors, even with the simplest equipment.
  
  "Before we get to the data, let me give some details about the
  measurement techniques.  The number of people entering the building
  is measured by a light beam and photocell across the door, connected
  to a standard PC.  An independent count is also provided by an
  ordinary mechanical turnstile right behind the door.  these devices
  were calibrated in my presence against a manual count; all three
  numbers agreed to within 1%.
  
  "The number of people leaving at the other end is measured by
  placing a temporay barrier in front of the exit door, which
  redirects the output flow to a fenced parking lot next to the
  building.  After the measurement period is over, the barrier is
  removed and the parking lot gate is closed.  The people in the
  enclosure are asked to stand still in rows, while Dr. Geller's
  assistants count them---twice, for good measure.
  
  "I can testify that the parking lot was empty before each run
  started. The fence around the parking lot is very tall and well
  guarded, so there is no question of random people from the street
  walking in and being included in the count.  Any errors that remain
  are definitely towards the conservative side; for instance, while
  Dr. Geller usually includes in the output count the few people who
  are still inside the machine by the end of the measuring period, I
  made a point of excluding them.
  
  I also didn't count several people who got carsick halfway through
  the ride, and were seen leaving the building through emergency exits
  in the back. (I verified that those exits are blocked by turnstiles
  that prevent people from entering through them.)  I also didn't
  correct our numbers to account for people who spontaneoulsy
  combusted or were abducted by gremlins, although I have reason to
  believe that about 20% of the people that enter are lost that way.
  Had I done so, the excess people count would have been over XXX%.
  
  "The only complication in carrying out these experiments is the
  narrow "operating window" of the machine.  Although the breeder can
  produce excess people reliably for years on end, it does so only if
  the rates at which people enter and exit the building are kept
  within rather tight limits.  Otherwise, no excess people are
  produced, and the breeder behaves just like an ordinary roller
  coaster.
  
  "The onset of the people-cloning regime is quite obvious, since it
  is always signalled by a sudden change in the quality of the noise
  made by the machine, and by a large and sudden drop ---20 to
  50%---in the input people flow rate.
  
  "In the first experiment I witnessed (Test 1), the cloning regime
  was not achieved, and accordingly my numbers showed no evidence of
  excess people. In fact, the output count I got was only XX% of the
  input.  This result confirms the claim that the measurement errors
  are conservative.
  
  "In Test 2, carred out that afternoon, the machine ran for a while
  in "roller coaster" mode, while Dr. Geller adjusted the input and
  output flows.  During that "warm-up" period the flow into the
  building was about XX person per minute, and any people who came out
  at the other end were allowed to walk away uncounted.  Afer about 20
  minutes, the machine suddenly went into "breeder" mode: its rythmic
  "tumtum-tumtum-tumtum" noise changed to a loud high-pitched whirr,
  while the input line slowed down to a trickle---about XX persons per
  minute.  At Dr. Geller command, the input counters were immediately
  reset and the exit line was diverted into the parking lot.
  
  "The measurements were prematurely ended after 19 minutes, when the
  machine shut down due to a power failure.  At that point the input
  counters were stopped, the parking lot was closed, and the people 
  in it were counted.  Here are the results for those 19 minutes:
  
     Output count: XXX people
     Input count:  XXX people (average XX people per minute).
     ------------------------
     Excess:       XXX people (everage XX people cloned per minute).
  
  "In Test 3, conducted the following day, the operating conditions of
  Test 2 were replicated as closely as possible, with similar results.
  I won't bother giving the numbers because you must all be pretty
  bored by now.
  
  "I must say that during my whole visit I didn't see a single
  pregnant woman entering the building; so there is no way that the
  excess people we measured was due to ordinary "biological" cloning.
  
  "I also looked carefully at the people coming out from the other
  end; and they definitely were real people, full-blooded and
  full-bodied.  In fact, the average size and weight of people in both
  lines was about the same; so it cannot be the case, as some have
  suggested, that the breeder works by splitting one fat lady into two
  circus midgets plus a refugee from Rwanda.
  
  "These experiments convinced me that Dr. Geller's Spinclone People
  Breeder does generate excess people in demographically significant
  amounts.  His very practical success fully vindicates the work of
  Drs. McFlons, Higgsman, Sukiyaki, Operetta, and hundreds of other
  honest scientists who have had to endure years of ridicule and abuse
  by the pet breeding establishment for daring to suggest that rabbits
  and white doves can be spontaneously generated inside top hats.
  
  (Credits: "Pirates of the Caribbean" must be a trademark of Walt
  Disney Productions of Hollywood, CA (or whatever); "Spinclone" is
  definitely a trademark of Industrial People Cloners, Washers and
  Driers of Palo Alto, CA.)