out so that the
diamond scratch now occupies a position somewhere near the centre of
this new capillary portion (Fig. 19, C).

4. Heat the tube in this position in the peep flame of the Bunsen
burner, and draw it out into a hair-like extremity. Snap off the glass
tube, leaving about 5 mm. of hair-like extremity attached to the upper
capillary portion (Fig. 19, D). Allow the glass to cool.

5. Lift up the bulb by the long capillary stem and allow the mercury to
return to its original position--an operation which will be facilitated
by snapping off the hair-like extremity from the long piece of capillary
tubing.

6. Mark on the capillary stem with a grease pencil the position of the
end of the column of mercury (Fig. 19, E.)

7. Warm the capillary tubing at this spot in the peep flame of the
Bunsen burner, and draw it out very slightly so that when cut at this
position a pointed extremity will be obtained.

8. With a glass-cutting knife cut the capillary tube through at the
point "b," and allow the mercury to run out.

9. Now apply a thick layer of sealing wax to the neck of the bulb.

10. Take a piece of 5 mm. bore glass tubing and draw it out as if making
an ordinary Pasteur pipette.

11. Break the capillary portion off so as to leave a covering tube
similar to that already used for the smaller graduated pipettes. Into
this covering tube drop the graduated bulb and draw the capillary stem
down through the conical extremity until further progress is stopped by
the layer of sealing wax.

12. Warm the pipette in the gas flame so as to melt the sealing wax and
make an air-tight joint.

13. Fit an india-rubber teat over the open end of the covering tube, and
the automatic pipette is ready for use (Fig. 19, F).

~Sedimentation Pipettes~ (Fig. 20).--These are prepared from 10 cm.
lengths of narrow glass tubing by sealing one extremity, blowing a
small bulb at the centre, and plugging the open end with cotton-wool;
after sterilisation the open end is provided with a short piece of
rubber tubing and a glass mouthpiece. When it is necessary to observe
sedimentation reactions in very small quantities of fluid, these tubes
will be found much more convenient than the 5 by 0.5 cm. test-tubes
previously mentioned.

[Illustration: FIG. 20.--Sedimentation pipette.]

Pasteur pipettes fitted with india-rubber teats will also be found
useful for sedimentation tests when dealing with minute quantities of
serum, etc.

[Il