286 Self riz.— Observations on some Physical Properties of 
Na 2 Si 0 3 together ( 35 , p. 681). This membrane may remain liquid for 
months. The duration of liquidity of most of these precipitation mem- 
branes ranges from 3 to 120 seconds. Quincke ( 36 , p. 722) describes 
these membranes as c a very thin invisible film of an oily liquid precipi- 
tate ’. In these physical, as well as in some of their osmotic properties, 
these precipitation membranes are very similar to the plasma-membrane. 
Whether or not the protoplasmic membrane is a precipitation membrane, 
one cannot say definitely. I am inclined to believe that it is not, primarily 
because of its ready reversibility. 
Thickness of the Plasma-membrane . As yet no mention has been 
made of the thickness of the plasma-membrane except to say that it is very 
delicate. Any attempt to estimate its thickness (after it is isolated and has 
coagulated) with the aid of a micrometer, the very graduating lines of 
which exceed in breadth the apparent thickness of the membrane, will, of 
course, be very crude. Yet I made such attempts and estimated the thick- 
ness to be less than one-fifth of a micron. This value is not far from that 
given by Ouincke for the thickness of some of the precipitation membranes 
with which he worked. He says ( 34 , p. 630), ‘ The thickness of this oil 
layer can, according to my investigation, be so small, less than 0-031 /x. that 
it can no longer be perceived with our best microscopes \ l 
Dewar ( 13 , p. 16), working on soap films, gives the far more minute 
thickness of soap bubbles ‘ thinned to the “ black ” stage ’ (this closely 
approaches the minimum thickness of such films) as ‘ 15 /x/x’s.’ Bechhold 
( 4 , p. 34), in reporting the work of others, also gives a very low figure. To 
quote : ‘ The thickness of the layer which will just form a solid skin has 
been measured, and found to be, for peptone 3 /x/x (Metcalf), for albumin 3 to 
7 /x/x (Devaux). Thus it is probably many times greater than the hypothetical 
diameter of a molecule, perhaps even equalling the radius of molecular 
attraction.’ That the value may not only equal but exceed the radius of 
molecular attraction is suggested by Quincke ( 35 , p. 631), who states that, 
1 I should like to call attention to some statements on the limit of size of objects visible and 
measurable with the modern microscope. Czapek ( 12 , p. 25) states that ‘ Ordinary microscopical 
observation with the strongest lenses can show particles of about 250 (x/x in diameter ’. One quarter 
of a micron (250///*) is not only not the limit of visibility of the strongest lenses, but is actually 
within the limit of fairly accurate mensurability. Quincke ( 34 , p. 630) remarks, as quoted above, 
that the thickness of a precipitation membrane is less than o-i /x, which cannot be perceived by the 
best microscopes. I cannot agree with either Quincke or Czapek, although Quincke’s statement is 
much nearer what I find to be true, namely, that the limit of visibility is about, but probably some- 
what less than, o-oooi mm. (o-i /x). The value given by Taylor ( 43 , p. 42) is somewhat above this. 
He says, ‘ A sol whose particles are less than about 0-15 [x will not be recognizable even with the 
best microscope (magnification 2,250)’. An extremely low value is given by Burton (6, p. 117) in 
a table of ‘ Lower limits of diameters of small particles ’. He gives the size of a particle visible 
under the ‘ordinary microscope’ as 2.5 x io -B cm. (0.25 /x/x is also given, but this is undoubtedly 
a typographical error, since 25-0 /xfx and not 0-25 [Xfx equals 2-5 x io“ 5 cm.). This value in micra is 
0-025 [x. Possibly Czapek had this figure in mind when he gave 250 fxfx as the limit, but misplaced 
the decimal point. We may safely conclude, then, that the limit of microscopical visibility is not 
above o-i fx. 
