FP00-0111-00 DESCRIPTION OPTICAL FIBER TECHNICAL FIELD This invention relates to an optical fiber which has 5 excellent transmission characteristics and can be used ideally in an optical fiber cable and an optical fiber coil. BACKGROUND ART An optical fiber has inside it a glass part made up of a core and a cladding, and has outside this cladding 10 a resin covering layer made up of one layer or a plurality of layers. It is well known that when an optical fiber suffers a side pressure it undergoes local bending, and this bending results in a worsening of transmission loss. Transmission loss caused by bending is called 'bend loss'. 15 There are two different types of bend loss: 'micro bend loss', which occurs at local minute bend parts resulting for example from adjacent optical fibers overlapping and exerting side pressures on each other; and 'macro bend loss ', which occurs as a result of an optical fiber itself describing 20 an arc when it is laid. If the radius of curvature of the arc described by an optical fiber itself is large, macro bend loss does not occur (or is small enough to be ignored). Because bend loss occurs at various locations on an optical transmission path, even if the loss at any one location 25 is not that large, over the whole optical transmission path it may become too large to be ignored. Consequently, optical 1 FPOO-0111-00 fibers which are resistant to bend loss are being sought after. And also in the case of optical fiber cables having optical fibers inside them, cables which are resistant to bend loss are being sought after. 5 Also, optical fiber coils, which are formed by winding an optical fiber into a coil, tend to have increased bend loss. An optical fiber coil is an optical component used in optical amplifiers and wavelength-dependent dispersion compensators, polarization mode dispersion compensators 10 and optical fiber gyros and the like. In optical fiber coils of this kind there is a tendency for the above-mentioned resin covering layer to be made thin in order to make the optical fiber in the coiled state compact, and in this case micro bending is more liable to occur. Consequently, in 15 optical fiber coils there is a tendency for worsening of transmission loss caused by bend loss to occur readily. A known optical fiber coil is disclosed for example in Japanese Patent Application Laid-open No. H10-123342. And specific products include for example Sumitomo Electric 20 Co., Ltd. coilsoftypenumbersDCFM-340, DCFM-680, DCFM-1020 andDCFM-1360. An optical fiber coil exerts a desired effect on an optical signal along its optical path. For example, an optical fiber coil used in an optical amplifier is an optical fiber doped with erbium (Erbium Doped optical Fiber, 25 or EDF) made into a coil, and amplifies an optical signal along the optical path of the optical fiber. 2 FPOO-0111-00 Here, because to amplify light an EDF of a certain length is necessary, the optical fiber is coiled so that it can be received efficiently inside the optical amplifier. The same applies to optical fiber coils used in other optical 5 components besides optical amplifiers, such as wavelength-dependent dispersion compensators, polarization mode dispersion compensators and optical fiber gyros. Optical fiber coils have generally been made by winding an optical fiber onto a bobbin. 10 However, in a wound optical fiber there remains tension, and this causes micro bend loss to occur. And also, because stress resulting from deformation of the bobbin acts on the optical fiber due to a difference in linear expansion coefficient between the bobbin and the optical fiber, the 15 transmission loss varies with temperature. To overcome this, as in the optical fiber coil disclosed in the above-mentioned application, various contrivances have been applied, and bobbinless optical fiber coils and bobbin structures withwhich the same effects as these can be obtained 20 have been studied. However, even with these various contrivances, it has not been possible to completely remove bend loss (micro bend loss) occurring as a result of minute bends in the optical fiber. And thus for optical fiber coils also, 25 improvements providing further enhancement of transmission characteristics have been sought after. 3 FPOO-0111-00 DISCLOSURE OF THE INVENTION An optical fiber provided by the present invention comprises a glass part having a core and at least one cladding and at least one covering layer formed around the glass 5 part, and the Young's modulus at 23 0 C of the covering layer, after the glass part is removed from it, is not greater than 400MPa. The optical fiber referred to here is not only a single core optical fiber and also includes multi-core optical fibers (for example tape-form optical fibers). 10 Here, preferably, the covering layer is made two layers, an inner layer covering and an outer layer covering, and the Young ' s modulus measured by the PIM method of this inner layer covering is made not greater than 0.8MPa. If this is done, transmission loss can be reduced by an effect of 15 the inner layer covering absorbing side pressures acting on the outer layer covering so that they do not reach the glass part. And, here, the sliding friction coefficient of the outermost layer of the covering layer is preferably made 20 not greater than 0.20. This sliding friction coefficient is a sliding friction coefficient relating to a covering layer of an optical fiber having inside it a glass part. The method by which the sliding friction coefficient is measured will be discussed later. 25 When this is done, contacting sites on adjacent optical fibers slide over each other readily; side pressure 4 FPO0-0111-00 components other than components normal to the surfaces of the optical fibers are lightened; and the side pressures that the optical fibers exert on each other decrease. As a result, with an optical fiber according to this invention, 5 transmission loss can be suppressed. Also, an optical fiber cable having the optical fiber described above inside it is desirable. The optical fiber cable referred to here is an optical fiber cable having inside it an optical fiber and having outside that a protective 10 covering sheath or the like. Inside the optical fiber cable, besides the optical fiber, there may also be provided a steel wire (high-tensile wire) for taking tensile forces acting on the optical fiber cable, and a spacer having a groove for receiving the optical fiber. The optical fiber 15 cable referred to here includes both cables having inside them a plurality of optical fibers and cables having only one optical fiber, and each optical fiber may be single-core or multi-core. Preferably, each optical fiber inside the optical fiber 20 cable is a dispersion-compensating optical fiber. The dispersion-compensating optical fiber (hereinafter also called a DCF) referred to here is an optical fiber having a wavelength-dependent dispersion characteristic of the opposite sign to an optical fiber for transmission path 25 use such as a single mode optical fiber, and is capable of canceling out wavelength-dependent dispersion in an 5 FPOO-0111-00 optical transmission path. Normally, it is modulized and used as a DCFM (Dispersion-Compensating optical Fiber Module). DCFMs are used in dispersion compensators and the like. 5 A dispersion compensator is a device for canceling out wavelength-dependent dispersion in the 1.55pm wavelength band in order to carry out long-distance, high-capacity optical transmission in the 1.55pm band using a single mode optical fiber having its zero dispersion wavelength in the 10 1.3pm band. This kind of dispersion compensator is made by making into a coil f or example by winding on a small-diameter bobbin a long DCF having a wavelength-dependent dispersion in the 1.55pm band of the opposite sign to that of a single mode optical fiber having its zero dispersion wavelength 15 in the 1.3pm band. Also, an optical fiber coil made by coiling the optical fiber described above is desirable. And preferably, in this optical fiber coil, the optical fiber is a dispersion-compensating optical fiber. An optical fiber 20 coil using a dispersion-compensating optical fiber is as discussed above. And in this kind of optical fiber coil, preferably, the coiled optical fiber is surrounded with a filler and the filler is a resin whose undisturbed penetration as 25 prescribed in JIS K 2220 is in the range of 5 to 200 over the entire measurement temperature range of -40 0 C to 100 0 C 6 FPO0-0111-00 and packs the whole coiled optical fiber and holds it in the coiled state. Undisturbed penetration is prescribed in Japanese Industrial Standards JIS K 2220-1993 (JIS K 2220-1993 2.(14), 5.3.1(4), 5.3.6 etc.). However, whereas 5 in JIS K 2220 the measurement temperature is 25 0 C, here, a resin whose undisturbed penetration is in the above-mentioned range over the entire measurement temperature range of -40*C to 100'C is used. BRIEF DESCRIPTION OF THE DRAWINGS 10 Fig. 1A is a cross-sectional view of an embodiment of an optical fiber according to the invention. Fig. 1B is a schematic view illustrating the refractive index distribution in a cross-section of an embodiment of an optical fiber according to the invention. 15 Fig. 2 is a perspective view of the optical fiber shown in Fig. 1. Fig. 3A is a perspective view illustrating a method for preparing a sample for Young's modulus measurement by the PIM method. 20 Fig. 3B is a perspective view of a sample for Young's modulus measurement by the PIM method. Fig. 4A is a schematic sectional view of the sample shown in Fig. 3B (before the application of a load) illustrating a method for obtaining a relationship between 25 load and displacement of when a load is applied to the glass part of any optical fiber in the sample. 7 FPOO-0111-00 Fig. 4B is a schematic sectional view of the sample shown in Fig. 3B (during the application of a load) illustrating a method for obtaining a relationship between load and displacement of when a load is applied to the glass 5 part of any optical fiber in the sample. Fig. 5A is a plan view of a jig used in measuring a sliding friction coefficient. Fig. 5B is a front view of a jig used in measuring a sliding friction coefficient. 10 Fig. 5C is a side view of a jig used in measuring a sliding friction coefficient. Fig. 6A is a perspective view of a bobbin (before an optical fiber is wound on it) used in an optical fiber coil. Fig. 6B is a perspective view of a bobbin (after an 15 optical fiber is wound on it) used in an optical fiber coil. Fig. 7 is a perspective view of an optical fiber cable. Fig. 8 is a plan view of an optical fiber module (with its cover opened). Fig. 9 is a sectional view of an optical fiber module. 20 BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of an optical fiber according to the invention will now be described, with reference to the drawings. Fig. 1A is a sectional view of the optical fiber 1 25 of this embodiment. The optical fiber 1 of this embodiment has inside it a glass part (a core 2 and a cladding 3) made 8 FPOO-0111-00 of vitreous silica. A resin covering layer 4 is formed on the outside of this glass part, and the resin covering layer 4 protects the glass part inside it. The cladding 3 is made up of a depressed cladding 3a 5 positioned immediately outside the core 2 and an outer layer cladding 3b outside the depressed cladding 3a. That is, the optical fiber 1 of this embodiment is a double-cladded DCF. The diameter of the core 2 is 2.7pm, the outer diameter of the depressed cladding 3a is 6.6pm, and the outer diameter 10 of the outer layer cladding 3b (that is, the outer diameter of the glass part) is 120pm. The resin covering layer 4 is made up of a primary covering layer 4a positioned immediately outside the cladding 3 and a secondary covering layer 4b positioned outside the 15 primary covering layer 4a. The thickness of the primary covering layer 4a is 10pm, and the thickness of the secondary covering layer 4b is 20pm. That is, the outer diameter of the secondary covering layer 4b (the outer diameter of the optical fiber 1) is 180pm. The primary covering layer 20 4a and the secondary covering layer 4b of the resin covering layer 4 on the optical fiber 1 are each made of a UV curable resin such as a urethane acrylate, and the primary covering layer 4a is made of a relatively soft resin while the secondary covering layer 4b is made of a relatively hard resin. 25 Fig. 1B shows the refractive index distribution of the optical fiber 1 of this embodiment. The distribution 9 FPOO-0111-00 graph shown in Fig. 1B shows the variation of the refractive index of the core 2 and the depressed cladding 3a with respect to the outer layer cladding 3b. As shown in Fig. 1B, the relative index difference (increase) A+ of the core 2 with 5 respect to the outer layer cladding 3b is 1.9%, and the relative index difference (decrease) A- of the depressed cladding 3a with respect to the outer layer cladding 3b is -0.4%. The wavelength-dependent dispersion in the 1.55pm 10 wavelength band of this optical fiber 1 is -120ps-nm- km- 1 , and the wavelength-dependent dispersion slope in the 1. 55pm wavelength band is -0.28ps nm-2km~ 1 . The wavelength-dependent dispersion slope shows the wavelength dependence of the wavelength-dependent dispersion value, 15 and is the gradient of the curve (or straight line) showing the wavelength-dependent dispersion value on a graph in which the horizontal axis is wavelength (nm) and the vertical axis is wavelength-dependent dispersion value (ps-nm~ 1 km- 1 ) . Also, the transmission loss in the 1.55pm wavelength band 20 is 0.40dB/km. Fig. 2 is a perspective view of this optical fiber 1. Also, when the Young's modulus at 23 0 C of the resin covering layer 4 of the optical fiber 1 of this embodiment was measured by a measurement method which will be discussed 25 hereinafter, it was not greater than 400MPa. The Young's modulus measurement method will be explained below. In 10 FPOO-0111-00 the following Young's measurement method, if an actual optical fiber is available, the Young's modulus of its covering layer can be measured without a sheet-form sample of the covering layer having to be obtained for the 5 measurement. First, the optical fiber the Young 's modulus of whose covering layer is to be measured is cut to a length of 8cm, and the cut optical fiber is put in a test tube containing a mixed solution (solvent) of acetone : ethanol = 7:3. 10 This test tube is then put in an ultrasonic cleaner containing water, for 30 seconds to 3 minutes. The length of this time is made the minimum length of time after which it is possible to remove the glass part of the optical fiber from inside the covering layer. This is because if even after 15 it has become possible to remove the glass part of the optical fiber from inside the covering layer the optical fiber is left in the ultrasonic cleaner for longer, the covering layer will break up. Next, the optical fiber is taken out of the test tube, 20 which has been taken out of the ultrasonic cleaner, and the glass part is removed from the covering layer. At this time, ample care is taken not to scratch the covering layer. If the covering layer becomes scratched, the procedure must be started again from the beginning. The covering 25 layer without the glass part is then held for 30 minutes in a thermostat vacuum tank held at pressure 10kPa (76mmHg) 11 FPOO-0111-00 or below, temperature 60 C, to evaporate the solvent. After that, the covering layer is held overnight in a desiccator, and then a tensile test is carried out with the following conditions. The conditions of the tensile test are: bench 5 mark (length on sample over which tensile force acts) = 25mm; pulling rate = 1mm/min; sample ambient temperature and sample temperature = 23*C; and sample ambient relative humidity 50% RH, and the result is obtained using the 2.5% secant formula. 10 This Young's modulus measurement method has the object of measuring the Young's modulus of the covering layer from an actual optical fiber. If the covering layer is a single layer, the Young ' s modulus obtained by the measurement method described above and the Young's modulus measured using a 15 sheet-form sample are approximately equal. When the covering layer has a multi-layer structure extending from an inner layer to an outer layer in the optical fiber, the Young ' s modulus obtained by the measurement method described above becomes approximately equal to the Young's modulus 20 of the outermost, hardest covering layer. The reason for this is as follows. Normally, the Young's modulus of an inner covering layer is made smaller than the Young's modulus of an outer covering layer. This is to reduce external forces acting 25 ontheglasspartoftheoptical fiberandtoprevent influences of temperature changes affecting the glass part. Because 12 FP0O-O111-00 of this, there is a difference of about 100 times between the Young's modulus of the inner covering layer and the Young's modulus of the outer covering layer, and the Young's modulus measured by the measurement method described above 5 reflects the Young's modulus of the outer, hard layer. That is, with the measurement method described above, if the covering layer is a single layer, the Young's modulus obtained is approximately equal to the sheet test value of the material of the covering, and if the covering layer 10 is multi-layer, it is approximately equal to the sheet test value of the material of the hardest, outer covering layer among the multiple layers. When by a measurement method described below the Young ' s modulus of the primary covering layer (inner layer covering) 15 4a of the optical fiber 1 of this embodiment was measured by the PIM method, it was not greater than 0.8MPa. The PIM Young's modulus measurement method will be explained below. In this case also, if an actual optical fiber is available, the Young's modulus of its covering layer can 20 be measured without a sheet-form sample of the covering layer having to be obtained for the measurement. Fig. 3A is a perspective view illustrating a method for preparing a sample for PIM-method Young's modulus measurement, and Fig. 3B is a perspective view of the sample. 25 Here, 17 is an optical fiber to be the object of measurement; 18 a pipe; and 19 the sample. The pipe 18 is made of acrylic 13 FPOO-0111-00 resin. Enough optical fibers 17 to substantially fill the pipe 18 are inserted into the pipe 18, and an adhesive consisting of an epoxy resin is poured into the gaps between the optical fibers 17 and between the inner wall of the 5 pipe 18 and the optical fibers 17, and hardened. Then, the pipe 18 is cut perpendicularly to its axis at two intermediate locations along the pipe 18, to make a disc-shaped sample 19. Both sides of the sample 19 are polished. The dimensions of the sample 19 are: thickness 10 L = 2mm; pipe inner diameter D1 = 4mm; pipe outer diameter D2 = 8mm. Fig. 4 is a sectional view illustrating a method for obtaining a relationship between load and displacement for any one of the optical fibers clustered in the sample 19, 15 Fig. 4A showing a state before a load is applied and Fig. 4B showing a state during load application. 20 is an optical fiber; 20a the glass part of the optical fiber; 20b an inner layer covering; 20c an outer layer covering; 21 a load-applying pin; and 21a the end face of the load-applying 20 pin 21. The load-applying pin 21 is shaped like an inverted truncated cone, and the diameter of its end face 21a is 50pm. The outer diameter Df of the glass part 20a of the optical fiber 20 is normally 100pm, and the outer diameter Dp of the covering 20b is normally 140pm. 25 With the cut and polished faces of the sample 19 kept horizontal, the end face 21a of the load-applying pin 21 14 FPOO-0111-00 is brought to abut with the upper end face of the glass part 20a of any of the optical fibers 20 in the sample 19 and a load is applied. Because the glass part 20a and the outer layer covering 20c are made of a material hard enough 5 for their deformation to be negligible compared to that of the inner layer covering 20b, and the amount of deformation caused to them by the applied load is amply small, only the inner layer covering 20b part deforms under the load applied with the load-applying pin 21. As a result, as 10 shown in Fig. 4B, the load-applying pin 21 and the glass part 20a displace downward. From the values of this loading and displacement of the load-applying pin 21, the Young's modulus of the inner layer covering is obtained using the following formula. 15 However, because when the load S is small the displacement Z is not proportional to the load S, the change AZ in the displacement Z is found in a place where the relationship between Z and S is substantially straight-line, for example in the range where the load S is between 5mN and 75mN (and 20 the change in load AS = 70mN), and this value is used. E = (1+.) AS x ln(Dp/Df)/(nLAZ) where E : Young's modulus of inner layer covering (N/mm 2 ) AS : change in load on load-applying pin (mN) 25 (in this case, 75mN-5mN=7OmN) Dp : outer diameter of inner layer covering (pm) 15 FPOO-0111-00 =140pm Df : outer diameter of optical fiber (pm) =100pm L : thickness of sample (mm) =2mm AZ : change in displacement of load-applying pin(pm) 5 . : Poisson's ratio (in this case, 0.45) When the sliding friction coefficient of the outermost layer (the secondary covering layer 4b) of the resin covering layer 4 of this embodiment was measured by a measurement method which will be discussed below, it was not greater 10 than 0.20. In this case also, if the following measurement method is used, as long as an actual optical fiber is available, this sliding friction coefficient can be measured without a sheet-form sample of the covering layer having to be obtained for the measurement. 15 A specific method for making the sliding friction coefficient of the secondary covering layer 4b less than 0.20 is to add a reactive silicone or a silicone oil such as silicone acrylate resin or the like to the secondary covering layer 4b to make its surface slippery. Or both 20 a reactive silicone and a silicone oil may be added. For example, by including 5% by weight of a silicone oil in the outermost covering layer of an ordinary optical fiber it is possible to bring its sliding friction coefficient to 0.18. This compares with a sliding friction coefficient 25 of 0.33 for an optical fiber containing no silicone oil (content 0%) in its outermost covering layer. Here, the 16 FP00-0111-00 content of the reactive silicone or silicone oil or the like is preferably at least 3%. When it is less than 3%, it is difficult to effectively make the covering layer surface slippery. 5 The method for measuring the sliding friction coefficient of the outermost covering layer will now be described. Figs. 5A through 5C are views illustrating a sliding friction coefficient measurement method, Fig. 5A being a top view, Fig. 5B a front view and Fig. 5C a side 10 view. 11 is a base member; 12 a base side fiber; 13 a slide member; 14 a slide side fiber; 15 a weight; and 16 a thread. The base member 11 has a horizontal, flat upper face and is made of a hard material such as a metal like aluminum or a hard plastic. The slide member 13 is a glass plate 15 of length 77mm x width 28mm x thickness 1mm and mass 4.9g. The weight 15 is of mass 50g. For the measurement, 60 fibers of length 20cm and 30 fibers of length 28mm are prepared from the optical fiber to be the object of measurement. Of these, the optical 20 fibers of length 20cm are divided into two groups of 30 and used as base side fibers 12. The fibers of length 28mm are divided into two groups of 15 and used as slide side fibers 14. The base side fibers 12 are arranged in parallel and closely within the groups on the upper face of the base 25 member 11 and affixed to the top of the base member 11 using double-sided adhesive tape or an adhesive. The spacing 17 FPOO-O111-00 between the two groups is made 10mm. The slide side fibers 14 are arranged in parallel without gaps within the groups with their length direction oriented in the width direction of the slide member 13 and affixed to one face of the slide 5 member 13 using double-sided adhesive tape or an adhesive. The two groups of slide side fibers 14 are disposed in positions 5mm away from the length direction ends of the slide member 13. Then, the slide member 13 is placed with its slide 10 side fibers 14 side downward on the base side fibers 12 affixed to the base member 11. At this time, the length direction of the slide side fibers 14 is kept perpendicular to the length direction of the base side fibers 12, and the length direction center of the slide side fibers 14 15 is kept between the two groups of base side fibers 12. The weight 15 is then placed on the middle of the upper face of the slide member 13, and a thread 16 for pulling the slide member 13 with is attached to the middle of one of the length direction ends of the slide member 13. 20 With the slide member 13 and the weight 15 placed on the base member 11 like this by way of the base side fibers 12 and the slide side fibers 14 disposed perpendicularly to each other, the thread 16 attached to the slide member 13 is pulled by a tensile tester (not shown) in the length 25 direction of the base side fibers 12, i.e. the direction of the arrow. The pulling rate at this time is made 60mm/min, 18 FPOO-0111-00 and the ambient temperature is kept to 23 0 C ±2 0 C. The value of the maximum tensile force (N) at this time divided by the weight corresponding to the sum of the masses of the slide member 13, the slide side fibers 14 and the weight 5 15, i.e. 0.54N, is obtained. Then, the slide side fibers and the base side fibers are replaced with new ones and the same measurement procedure is carried out again. This measurement procedure is repeated until it has been carried out 10 times, and the average 10 of the ten values of the maximum tensile force (N) / total mass (0.54N) thus obtained is taken as the sliding friction coefficient of the outermost covering layer of the object optical fiber. That is, the average of the maximum static friction coefficients obtained in a test using the test 15 apparatus described above is the sliding friction coefficient. Because the Young's modulus explained above of the covering layer of the optical fiber 1 is made not greater than 400MPa, when the optical fiber 1 is bent, the force 20 exerted by the resin covering layer 4 on the glass part consisting of the core 2 and the cladding 3 is lightened, and bend loss is reduced. When the Young's modulus exceeds 400MPa, this 'bend loss is not reduced. And because in the optical fiber 1 of this embodiment 25 the sliding friction coefficient of the outermost layer (the secondary covering layer 4b) of the resin covering 19 FP00-0111-00 layer 4 is made not greater than 0.20, contacting sites on adjacent optical fibers 1 slide over each other readily; side pressure components other than components normal to the surfaces of the optical fibers 1 are lightened; and 5 the side pressures that the optical fibers 1 exert on each other decrease. As a result, with the optical fiber 1, micro bend loss caused by side pressures is reduced and transmission loss can be suppressed. When the sliding friction coefficient exceeds 0.20, because contacting sites 10 on adjacent optical fibers 1 do not slide over each other, side pressures tend to act at those sites, micro bend loss occurs and transmission loss is not reduced. To check the effect of the Young' s modulus of the resin covering layer 4 being made not greater than 400MPa and 15 the effect of the sliding friction coefficient of the outermost layer (the secondary covering layer 4b) being made not greater than 0.20, an optical fiber coil was made using the optical f iber 1 described above and its transmission loss was measured. The optical fiber was made into an optical 20 fiber coil for the measurement because coiling an optical fiber tends to cause bend loss such as micro bend loss and macro bend loss to occur. Here the measurement was carried out after the optical fiber 1 was made into an optical fiber coil. A 5km length 25 of the optical fiber 1 was wound onto an aluminum bobbin 8 having a fiber winding part 8a of the kind shown in Fig. 20 FPOO-0111-00 6A to obtain the optical fiber coil 9 shown in Fig. 6B, after which the bobbin 8 was removed to leave a coil state. Signallightof wavelength 1. 6pmwas used for the measurement, and the transmission loss at room temperature (23 0 C) and 5 at -20 0 C was measured. The outer diameter of the fiber winding part 8a was 120mm. The optical fibers used in the test had the different Young' s modulus of the resin covering layer 4 and the different sliding friction coefficients of the outermost layer (the 10 secondary covering layer 4b) shown in [ Table 1] . Embodiments 1 through 3 in [Table 1] are the optical fiber 1 of the present invention, and for comparison with these the test was also carried out on Comparison Examples 1 and 2. The test results are also shown in [Table 1]. 15 [Table 1] Young's sliding Friction Transmission Loss (dB/km) Modulus Coefficient RoomTemp. (MPa) (23 0 C) -200C Embodiment 1 400 0.18 0.33 0.38 Embodiment 2 200 0.18 0.33 0.35 Embodiment 3 400 0.25 0.35 0.39 Comp. Ex. 1 800 0.25 0.50 1.0 Comp. Ex. 2 800 0.18 0.33 0.83 As is clear from the results shown in [Table 1], it was found that the transmission characteristics are superior 20 when the Young's modulus of the resin covering layer 4 is 21 FPOO-0111-00 400MPa or less (Embodiment 3 and Comparison Example 1; Embodiments 1, 2 and Comparison Example 2). Also, it was found that the transmission characteristics are superior when the sliding friction coefficient of the outermost 5 covering layer (the secondary covering layer 4b) is 0.20 or less (Embodiment 1 and Comparison Example 3; Comparison Example 1 and Comparison Example 2). Of course, it is preferable if the Young's modulus of the resin covering layer 4 is 400MPa or below and also the sliding friction 10 coefficient of the outermost covering layer (the secondary covering layer 4b) is 0.20 or below, and this is the case in which the best transmission loss is obtained (Embodiments 1, 2). To investigate the effect of only making the sliding 15 friction coefficient of the outermost layer (the secondary covering layer 4b) of the resin covering layer 4 not greater than 0.20 (with the Young's modulus of the resin covering layer 4 not necessarily being 400MPa or below), the following measurement was also carried out. The optical fiber used 20 here was a DCF, and for the measurement it was made into a DCF coil bundle. The cross-section of this DCF is the same as that shown in Fig. 1, but its dimensions were different from that described above. The outer diameter of the optical fiber 25 1 (i.e. the outer diameter of the secondary covering layer 4b) was 180pm, the outer diameter of the primary covering 22 FPOO-0111-00 layer 4a was 140pm, and the outer diameter of the glass part (i.e. the outer diameter of the cladding 3) was 100pm. The layer thicknesses of the primary covering layer 4a and the secondary covering layer 4b were each 20pm. The 5 outer diameter of the core 2 was 2.65pm, the outer diameter of the depressed cladding 3a was 7 .58pm, and the outer diameter of the outer layer cladding 3b was 100pm. The relative index difference (increase) A+ of the core 2 with respect to the outer layer cladding 3b was 2.1%, and the relative 10 index difference (decrease) A- of the depressed cladding 3a with respect to the outer layer cladding 3b was -0.4%. At wavelength 1. 55pm, the wavelength-dependent dispersion and the wavelength-dependent dispersion slope of this optical fiber were respectively -100ps nm~ km- 1 and -0.29ps-nm 2 .km- 1 . 15 The covering on the optical fiber was two-layered, and the materials of the inner layer covering and the outer layer covering were as shown in [Table 2] below. The covering resin material types were as follows: UV resin a: relatively soft urethane acrylate resin 20 (Young' s modulus by PIMmethod: 1.2MPa) UV resin b: urethane acrylate resin softer still than UV resin a (Young's modulus by PIM method: 0.7MPa) UV resin A: relatively hard urethane acrylate resin with 25 silicone acrylate resin added UV resin B: relatively hard urethane acrylate resin with 23 FPOO-0111-00 silicone oil added UV resin C: relatively hard urethane acrylate resin (no silicone acrylate resin or silicone oil added) 5 [Table 2] Covering Material Sliding Friction Transmission Inner Coefficient of Coil State Loss Layer Outer Layer Outer Layer Increase Covering Covering Covering Surf ace (dB/km) Ref. Ex. 1 UV resin a UV resin A 0.18 bobbin 0.08 Ref. Ex. 2 UV resin a UV resin A 0.18 no foin; 0.03 no bin; 0.02 Ref. Ex. 3 UV resin b UV resin A 0.18 no f n; 0.0 nobobbin 0.3 Ref. Ex. 4 UV resin a UV resin B 0.18 no bobbin 0.1 Comp. Ex. 3 UV resin a UV resin C 0.25 bobbin 0.35 Comp. Ex. 4 1UV resin a UV resin C 0.25 no bobbin 0.15 A 250pm thick sheet of the relatively hard urethane acrylate resin mentioned above was hardened by ultraviolet 10 irradiation and a no. 2 dumbbell sample piece was taken from it and its Young's modulus measured in accordance with JIS K 7113, with a pulling rate of 1mm/min, and found to be 700MPa. For each of the optical fibers made with different 15 types of covering resin the sliding friction coefficient of the outermost layer (the secondary covering layer 4b) was measured by the method described above, and the results shown in [Table 2] above were obtained. According to these results, in all of the Reference Examples 1 through 4 the 24 FP00-0111-00 sliding friction coefficient was less than 0.20. And in Comparison Examples 3 and 4 the sliding friction coefficient was large, at 0.25. Next, a 5km length of the optical fiber was wound into 5 a coil on a bobbin of core diameter 120mm at a winding tension of 0.4N. The coil state in each case was as shown in [Table 2] from among the following three: with bobbin : wound on a bobbin no bobbin : wound on a bobbin and then removed from 10 the bobbin (the bundled state of the coil in Comparison Example 4 did not disintegrate much after the coil was removed from the bobbin) no bobbin; filler : wound on a bobbin, removed from the bobbin, and then packed in a silicone gel-form filler 15 (the coils of Reference Examples 2 through 4 disintegrated slightly when removed from the bobbin) When for each of the finished Reference Examples 1 through 4 and Comparison Examples 3 and 4 the transmission loss increase at wavelength 1.6pm from before the optical 20 fiber was wound on the bobbin of core diameter 120mm was measured, the results shown in [Table 2] were obtained. Here the respective transmission loss increases at wavelength 1.6pm were compared to facilitate comparison, because even when the transmission loss increase at wavelength 1.55pm 25 is very small the transmission loss increase at wavelength 1.6pm is normally relatively large. 25 FPOO-0111-00 From a comparison of Reference Examples 1 through 4 and Comparison Examples 3 and 4 it can be seen that the transmission loss increase at wavelength 1.6pm is smaller when the sliding friction coefficient of the outermost layer 5 (the secondary covering layer 4b) is lower. And if the friction coefficient is 0.18, even if the coil is still wound on the bobbin, the transmission loss increase at wavelength 1. 6pmcan be kept to a smallvalue of about 0. 08dB/km. And from a comparison of Reference Example 1 and Comparison 10 Example 2 it can be seen that the transmission loss increase is lower when after the optical fiber is wound on the bobbin the coil is removed from the bobbin before being packed in the filler than when it is packed in the filler while still on the bobbin. 15 From a comparison of Reference Example 2 and Reference Example 3 it can be seen that if the inner layer covering (the primary covering layer 4a) is made an even softer one the transmission loss increase can be suppressed further. And from a comparison of Reference Example 2 and Reference 20 Example 4 it can be seen that the addition of silicone acrylate resin and the addition of a silicone oil to the material of the outer layer covering (the secondary covering layer 4b) each have the same level of effect of suppressing transmission loss increase. 25 26 FP0O-0111-00 [Table 3] Sliding Friction Transmission Loss (dB/km) Coefficient of Outer Layer RoomTemp. -20 0 C Covering Surface (23*C) Ref. Ex. 5 0.18 0.33 0.40 Ref. Ex. 6 0.17 0.33 0.39 Comp. Ex. 5 0.25 0.50 0.53 To further investigate the effect of only making the sliding friction coefficient of the outermost layer (the 5 secondary covering layer 4b) of the resin covering layer 4 not greater than 0.20 (with the Young's modulus of the resin covering layer 4 not necessarily being 400MPa or below) , the following measurement was also carried out. Again, the optical fiber used for the measurement was a DCF, and 10 although its cross-section was the same as that shown in Fig. 1 the dimensions of its parts were different. The diameter of the core 2 was 2.7pm, the outer diameter of the depressed cladding 3a was 6 .6pm, and the outer diameter of the outer layer cladding 3b (i.e. the outer diameter 15 of the glass part) was 120pm. The thickness of the primary covering layer 4a was 40pm, and the thickness of the secondary covering layer 4b was 20pm. That is, the outer diameter of the optical fiber 1 (the outer diameter of the secondary covering layer 4b) was 240pm. 20 The refractive index distribution of this optical fiber 1 was similar to that shown in Fig. 2, but the relative 27 FP0O-0111-00 index difference (increase) A+ of the core 2 with respect to the outer layer cladding 3b was 1.9%. The relative index difference (decrease) A- of the depressed cladding 3a with respect to the outer layer cladding 3b was -0.4%. At 20 0 C 5 the Young's modulus of the primary covering layer 4a was 0.98MPa and the Young's modulus of the secondary covering layer 4b was 784MPa (the Young's modulus here was measured using a sheet sample). The wavelength-dependent dispersion of this optical 10 fiber 1 in the 1.55pm wavelength band was -120ps-nm- km~ 1 , and the wavelength-dependent dispersion slope in the 1. 55pm wavelength band was -0 .28ps-nm- km- 1 . The transmission loss in the 1.55pm wavelength band was 0.33dB/km. The sliding friction coefficient of the outermost layer 15 (the secondary covering layer 4a) of the optical fiber 1 described above was varied and an experiment was carried out. In the experiment, 5km of the optical fiber with its sliding friction coefficient varied was wound on a bobbin of diameter 120mm and then the bobbin was removed and the 20 transmission loss of the coiled optical fiber was measured. The signal light used at the time of the measurement was 1.6pm, and the temperature at the time of measurement was room temperature (23 0 C) and -20 0 C. The results of the experiment are shown in [Table 3]. 25 As can be seen from [Table 3], when the sliding friction coefficient of the outermost layer (the secondary covering 28 FPOO-0111-00 layer 4a) is not greater than 0.20, transmission loss can be kept low. Next, an optical fiber cable using the optical fiber discussed above will be described. This optical fiber cable 5 10 is shown in Fig. 7. The optical fiber cable 10 has inside it the optical fiber 1 described above. The optical fiber cable 10 of this embodiment is a cable having one single core optical fiber 1 of the kind described above and is also known as an optical fiber cord. 10 The optical fiber cable 10 has at its center the optical f iber 1 described above, and around that is formed a cushioning layer 10a. The cushioning layer 10a is made of fibers or the like, and on the outside of this is a covering layer 10b made out of a synthetic resin. Although an optical 15 fiber cable having inside it only a single optical fiber is being taken as an example here, such cables sometimes have a steel wire (high-tensile wire) for taking tensile forces acting on the optical fiber cable and a spacer having a groove for receiving the optical fiber. And the optical 20 fiber cable may have inside it a plurality of optical fibers, or an optical fiber inside it may be a multi-core optical fiber (a tape-form optical fiber, or optical fiber ribbon). An optical fiber cable may with its covering sheath or steel wire exert side pressures and distortion on the 25 optical fiber. And when there are multiple optical fibers, the optical fibers may exert side pressures on each other. 29 FPOO-0111-00 However, because as described above the optical fiber 1 does not readily suffer transmission loss increases caused by side pressures and has superior transmission characteristics, the optical fiber cable 10 of this 5 embodiment can also have excellent transmission characteristics. Next, an optical fiber coil using the optical fiber 1 discussed above will be described. As shown in Fig. 8 and Fig. 9, as this optical fiber coil, a coiled optical 10 fiber 1 whose winding distortion has essentially been released is received in a receiving case 80 having a rectangular bottom. The ends of the optical fiber 1 are connected by fused parts 44 to pigtail fibers 45. The inside of the receiving case 80 is filled with a filler 84, which 15 packs in the optical fiber 1. A cover 82 is fitted to the receiving case 80 and sealed. The coiled optical fiber 1 is embedded within the filler 84. In addition to the optical fiber 1 itself being resistant to transmission loss increases caused by side pressures, 20 transmission loss increases are also prevented by the optical fiber coil having being removed from the bobbin. Because it has been removed from the bobbin, micro bends forming as a result of the optical fiber 1 being pressed against the bobbin under the winding tension or by adjacent winds 25 of the optical fiber 1 exerting side pressures on each other also tend to be eliminated. And because its coiled state 30 FPOO-0111-00 is held by the filler 84, changes in transmission characteristics caused by uncoiling of the optical fiber 1 do not occur either. As shown in Fig. 9, the filler 84 is made up of a fluid 5 part 84a and an elastic solid part 84b. The part making contact with the optical fiber 1 is the fluid part 84a, which is liquid or semi-liquid, and outside of this is the elastic solid part 84b, which has some elasticity. Here, liquid or semi-liquid refers to a state such that the optical 10 fiber 1 can change its position freely inside the fluid part 84a and such that after it changes its position it essentially is not subjected to a force from the liquid or semi-liquid part tending to return it to the original position. 15 Here, the fluid part 84a and the elastic solid part 84b are made of the same material, but their hardnesses differ. In the manufacture of the optical fiber coil described above, the coiled optical fiber 1 is embedded in a liquid or semi-liquid filler 84 in a receiving case 20 80, and the filler 84 is then hardened. or the coiled optical fiber 1 is placed inside the receiving case 80 first and then the receiving case 80 is filled with the filler 84, and the filler 84 is then hardened. Here, the fluid part 84a and the elastic solid part 25 84b described above are formed by the hardness around the optical fiber 1 being made lower than that elsewhere. 31 FP00-0111-00 Although in Fig. 9, for convenience, the boundary between the fluid part 84a and the elastic solid part 84b has been made clear, in this embodiment the boundary between the fluid part 84a and the elastic solid part 84b is not a clearly 5 formed boundary but rather is a boundary at which the hardness changes gradually. The wavelength-dependent dispersion value and the wavelength-dependent dispersion slope value of the optical fiber 1 are of the opposite sign to those of an optical 10 fiber for transmission path use to be connected to this optical fiber coil. That is, by using this optical fiber coil, it is possible to compensate for wavelength-dependent dispersion arising in an optical fiber of a transmission path. And for wavelength band transmission use, by making 15 the wavelength-dependent dispersion slope the opposite sign, it is possible to compensate for wavelength-dependent dispersion over the wavelength band. As the filler 84, a silicone resin hardened by heat or ultraviolet radiation, or a highly viscous jelly-form 20 admixture made by swelling a butadiene or silicone rubber with a solvent such as silicone or naphthene and adding another resin or the like as necessary, or an oil such as silicone oil can be used. As mentioned above, here, the fluid part 84a and the elastic solid part 84b are formed 25 by changing the hardness of the filler 84. A method for changing the hardness of the filler 84 will now be described. 32 FPOO-0111-00 Platinum is included in the filler 84 as a catalyst for a hardening reaction. The functioning of this platinum catalyst is inhibited by an isocyanate compound. That is, an isocyanate compound can be used as a hardening inhibitor. 5 Here, this isocyanate compound is included in the secondary covering layer 4b on the outside of the optical fiber 1. The isocyanate compound may be included not only in the secondary covering layer 4b of the optical fiber 1 but also in the primary covering layer 4a. 10 In this way, hardening of the parts of the filler 84 in contact with the optical fiber 1 is inhibited by the isocyanate compound included in the secondary covering layer 4b, and the fluid part 84a is formed. Parts more than a certain distance away from the optical fiber 1 are not 15 prevented from hardening by the isocyanate compound included in the secondary covering layer 4b and harden normally, whereby the elastic solid part 84b is formed. If in this way the same starting material is used for both the fluid part 84a and the elastic solid part 84b, and the fluid part 20 84a and the elastic solid part 84b are formed by varying its hardness, the manufacture of an optical fiber coil can be carried out more efficiently. And, here, including a hardening inhibitor in the secondary covering layer 4b of the optical fiber 1 makes 25 it easy to reduce the hardness of only the parts where the fluid part 84a is to be formed, and is very convenient in 33 FPOO-0111-00 the manufacture of the optical fiber coil. And, at this time, platinum is included in the filler 84 as a hardening reaction catalyst, and an isocyanate compound is used as the hardening inhibitor. The combination of a hardening 5 reaction catalyst and a hardening inhibitor for canceling out the action of the catalyst like this utilizes the process of hardening the filler 84 of the optical fiber coil effectively and makes it possible to form the fluid part 84a well. 10 When the coiled optical fiber 1 is fixed with an ordinary adhesive or resin, because the Young's modulus of the resin when hardened reaches over 50OMPa, excessive pressing forces act on the optical fiber 1 and undesirable bending distortion occurs. By forming a fluid part 84a around the optical 15 fiber 1 like this, on the other hand, it is possible to fix the optical fiber surely without exerting excessive pressing forces which would produce bending distortion on the optical fiber 1 constituting the optical fiber coil. And, the optical fiber 1 is not rigidly restrained 20 inside the fluid part 84a. Consequently, even if the thermal expansivities of the optical fiber 1 and the filler 84 (the elastic solid part 84b) are different, the optical fiber 1 can to a certain extent change its position inside the fluid part 84a, and temperature changes are prevented from 25 resulting in external forces acting on the optical fiber 1. That is, the optical fiber coil also has superior 34 FP 00-0111-00 temperature characteristics. To investigate the effect of only holding the optical fiber with a filler forming a liquid part or a semi-liquid part around the optical fiber (with the Young's modulus of the resin covering layer 4 5 not necessarily being 400MPa or below), a 5-cycle, -20 0 C to 70 0 C, 6-hour hold at each, thermal cycle test was carried out. The optical fiber 1 used here was a double-cladded DCF having the same cross-section as that shown in Fig. 10 1A, the diameter of the core 2 being 2.7pm and the diameter of the depressed cladding 3a being 6. 6pm. The outer diameter of the glass part was 120pm, the thickness d of the primary covering layer 4a was 15pm, the thickness e of the secondary covering layer 4b was 15pm and the outer diameter f of the 15 optical fiber l as a whole was 180pm. A+ and A-, respectively the deviations in the refractive indexes of the core 2 and the depressed cladding 3a with respect to the refractive index of the outer layer cladding 3b, were 1.9% and -0.4%. The wavelength-dependent dispersion and 20 wavelength-dependent dispersion slope of this DCF were respectively -120ps nm- km~ 1 and -0.28ps nm-2.km- 1 at wavelength 1.55pm, and the transmission loss was 0.40dB/km. However, in the optical fiber 1 used here, the Young's modulus of the covering layer 4 was not necessarily 400MPa 25 or less and the Young's modulus of the primary covering layer 4a measured by the PIM method was not necessarily 35 FP 0-0111-00 0.8MPa or less. The difference between the maximum value and the minimum value among ten loss measurement values, five loss values measured at -20 0 C and five loss values measured at 70 0 C, 5 was calculated as a level of loss fluctuation. The result was that the loss fluctuation level at measurement wavelength 1.6pm was 0.ldB/km. To make a comparison with the optical fiber coil described above, the same measurement was carried out using an optical fiber coil wherein no fluid part 84a 10 was formed around the optical fiber 1 and instead the filler 84 was hardened completely. The result of this was that the loss fluctuation level at the measurement wavelength 1.6pm was 1.2dB/km. Thus it was found that, as is clear from this result, an optical fiber coil having a filler 15 forming a liquid part or a semi-liquid part of the kind described above has extremely good temperature characteristics. The elastic solid part 84b after hardening is a material whose undisturbed penetration as prescribed by JIS K 2220 20 is in the range of 5 to 200 over the entire temperature measurement range of -40 0 C to 100 0 C. The temperature range of -40 0 C to 100*C is the range of temperatures over which the optical fiber coil is actually used. Although by forming the fluid part 84a it is possible to prevent excessive forces 25 fromactingontheoptical fiber 1, there is stillapossibility of the optical fiber 1 making contact with the elastic solid 36 FP0O-0111-00 part 84b; however, by giving the elastic solid part 84b the above-mentioned property, external forces can be better prevented from loading the optical fiber 1. When the undisturbed penetration of the elastic solid 5 part 84b is less than 5, losses at high wavelengths become too large due to micro-bending of the optical fiber 1, and it is of no practical use. And when the undisturbed penetration exceeds 200, because the elastic solid part 84b cannot hold the shape of the fluid part 84a and the 10 coiled state of the optical fiber 1 inside the fluid part 84a, during use the fluid part 84a ceases to be maintained or the coil uncoils, and the transmission characteristics cannot be stabilized. The receiving case may alternatively be doughnut-shaped 15 in form and its bottom may be curved instead of flat. And if a liquid part is formed, the method by which it is manufactured does not have to be the manufacturing method of the embodiment described above. For example, the whole of the coiled optical fiber may be dipped in oil to coat 20 the optical fiber with oil and then this may be molded with a filler in the normal way. or there is also the method of adjustingthemixproportions of abaseagentandahardening agent of the filler to facilitate the making of an unhardened state to constitute a liquid part. 25 And as the form of the liquid part also, just the close proximity of the optical fiber may be the liquid part, as 37 FP00-0111-00 described above, or alternatively a single, large, doughnut-shaped liquid part in which the whole coiled optical fiber can be received may be formed. An optical fiber coil using a filler having the liquid part or semi-liquid part 5 described above can improve the transmission characteristics of any kind of optical fiber. This invention is not limited to the embodiments described above. For example, the optical fiber 1 is not limited to an optical fiber having a double cladding type 10 refractive index distribution. It can also be an optical fiber having a refractive index distribution called matched cladding type, double core type or segment core type. And a structure wherein the refractive index of a groove part between a central core and a segment core is made smaller 15 than that of a cladding part can be employed. And whereas in the foregoing description the dispersion-compensating optical fiber was taken to be one having a negative wavelength-dependent dispersion at wavelength 1.55pm, when the wavelength-dependent dispersion in the 1. 55pm wavelength 20 band of a single mode optical fiber is negative, the wavelength-dependent dispersion of the dispersion-compensating optical fiber should be made the opposite sign, i.e. a positive wavelength-dependent dispersion. 25 Also, although the optical fiber 1 described above was a dispersion-compensating optical fiber (DCF), it may 38 FP00-0111-00 be some other type of optical fiber. For example, it may be a single mode optical fiber or a wavelength-dependent dispersion-shifted optical fiber, an NZ wavelength-dependent dispersion-shifted optical fiber, an 5 erbium-doped optical fiber, or a polarization-maintaining optical fiber. Using these kinds of optical fiber also, an optical fiber coil having superior transmission characteristics can be obtained. The single mode optical fiber (SMF) is an optical fiber 10 designed for the main purpose of transmitting optical signals in the 1.3pm wavelength band. When this optical fiber is used to transmit optical signals in the 1.55pm wavelength band, the phenomenon of wavelength-dependent dispersion arises. This wavelength-dependent dispersion is 15 compensated for with the above-mentioned DCFM or the like. Reversely, an SMF is also used for compensating with its own positive wavelength-dependent dispersion for an optical signal having a negative wavelength-dependent dispersion froma DCF or the like. In this case, it is sometimes modulized 20 for use. The dispersion-shifted optical fiber (or DSF) is an optical fiber designed for the main purpose of transmitting optical signals in the 1.55pm wavelength band. It has the characteristic that its wavelength-dependent dispersion 25 value in the 1.55pm wavelength band is zero, but when transmitting with wavelength division multiplexing 39 FP00-0111-00 (hereinafter also referred to as WDM) it suffers the wavelength-dependent dispersion phenomenon on the long and short wavelength sides of this wavelength band. A DSF is sometimes used as an optical fiber for Raman scattering 5 excitation. It is sometimes modulized for use. The NZ (Non-Zero) dispersion-shifted optical fiber (or NZ-DSF) is an optical fiber designed with the center wavelength at which its wavelength-dependent dispersion is zero somewhat shifted from 1.55pm to reduce nonlinear 10 effects arising in the case of the above-mentioned DSF. It is the same as the above-mentioned DSF in that the wavelength-dependent dispersion phenomenon occurs in transmission with WDM. It is sometimes modulized for use. The erbium-doped optical fiber (or EDF) is an optical 15 fiber whose core has been doped with erbium ions. When while absorbing light of wavelength 0.98pm or 1.48pm it is fed with 1. 55pm wavelength band signal light it provides induced emission and increases the power of the signal light. Normally it is used in a modulized state as an optical 20 amplifier (Erbium-Doped optical Fiber Amplifier, or EDFA) or the like. The polarization-maintaining optical fiber (PMF) is an optical fiber which transmits linearly polarized light with its polarization maintained, and is used in optical 25 fiber gyros and polarization mode dispersion compensators and the like. It is normally used modulized as a PMFM 40 FPO0-0111-00 (Polarization-Maintaining optical Fiber Module). INDUSTRIAL APPLICABILITY In an optical fiber according to this invention, because the Young' s modulus of the covering layer is made not greater 5 than 400MPa, side forces acting on the optical fiber are moderated by the covering layer and their transfer to the glass part is lessened. As a result, it is possible to prevent worsening of transmission loss caused by side pressures and realize superior transmission loss 10 performance. Here, if the covering of the optical fiber is made double-layer and the Young's modulus of the inner layer covering measured by the PIM method is made not greater than 0.8MPa, because side pressures acting on the outer 15 layer covering are absorbed by the inner layer covering and are not readily transmitted to the glass part, increases in transmission loss caused by coiling can be further reduced. And, when the sliding friction coefficient of the outermost covering layer is made not more than 0.20, adjacent 20 optical fibers slide easily at their contacting surfaces and the external forces they exert on each other are reduced. By this means, it is possible to reduce forces acting on the glass part inside the optical fiber and realize still more superior transmission characteristics. 25 41