It is expected that the transverse profile of the beam will not be Gaussian distribution, but will be accompanied by a long tail based on experience involving SLC experiments.  While the origin of the tail is not well known at present, we will conservatively assume that the beam has a flat tail beyond in the horizontal(x) and vertical(y) directions with a relative intensity of 1%. As can be clearly seen in the next section, the beam has to be collimated within and in order to control the background of the synchrotron radiation. Since the real size of the beam is on the order of m, only the tail must be expanded enough to avoid any destructive deformation of the beam by an interaction between the Gaussian core and a collimator, which is non-linear collimation method.  This is one reason why a 1200m long collimator section is necessary for a 250GeV beam. One RF pulse can accelerate a train of 8550 bunches separated by 1.45.6nsec, each at 150Hz. Since each bunch consists of electrons (or positrons), (1%tail)(bunches) electrons will hit collimators at 150Hz.
Figure 4: Muon spoiler for NLC(SLAC) presented by L.Keller at LC93. A square tunnel of 3x3m is filled by the 1.5 Tesla magnetized spoiler, the length and weight of which are 9.1m and 750 tons, respectively.
Figure 3: Muon attenuator for JLC-I(KEK) presented by Y.Namito at LC93. Two iron pipes are magnetized axially in opposite directions for both charged muons which can be trapped; also, the 120m length of the iron pipe corresponds to a mean range of 250GeV muons.
Two detailed simulations of the muon background has been made for JLC-I and NLC by Y.Namito and L.Keller,  respectively. Y.Namito employed a new method for a muon attenuator, which was originally proposed by E.A.Kushnirenko.  The principle idea is to confine and inside two iron pipes magnetized axially in opposite directions and to absorb muons by the energy loss, as shown in Fig.3. There are four sets of collimators accompanied by a muon attenuator in the collimator section in order to collimate the horizontal and vertical beam profiles twice for a smearing probability. The best attenuation was obtained to be with zero magnetic field in a smaller tunnel(R=1.75m) at the furthest collimation point. For NLC, L.Keller optimized a system of three and half muon spoilers filling a square tunnel of 3x3m(Fig.4) based on valuable experience concerning SLC experiments, where the half-filling spoiler is located inside a big bend, and the three are in a final focus system. The result was better than the case of the JLC-I; that is, it can tolerate electrons loss per train corresponding to . Though more studies are needed concerning cost performance and engineering work, the optimum design would be a combination of attenuators and the spoilers located at the collimation section and the final focus system, respectively.