Further Developments in Scavenging Analysis for Two-Cycle Engines

Author(s): G. P. Blair and R. G. Kenny

Source: SAE Transactions, Vol. 89, Section 1: 800004–800252 (1980), pp. 231-245
Published by: SAE International
Stable URL: https://www.jstor.org/stable/44632419
Accessed: 02-10-2021 05:27 UTC

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/44632419?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

SAE International is collaborating with JSTOR to digitize, preserve and extend access to SAE
Transactions

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 800038

Further Developments in
Scavenging Analysis for
Two-Cycle Engines
G. P. Blair and R. G. Kenny
Dept. of Mechanical and Industrial Engrg.,
The Queen's Univ. of Belfast (Northern Ireland)

MUCH HAS BEEN WRITTEN ON THE SUBJECT in a general sense for all two-cycle engines.
OF SCAVENGE FLOW in two-stroke cycle engines, Doubtless many of the actual engines used in
be they naturally aspirated, crankcase com- the several papers were considerably improved
pression spark-ignited units or turbocharged in the scavenging flow sense by the experi-
Diesel engines. An earlier paper (1)* by the mental effort expended. One paper (9) given
author briefly reviewed the literature in the by Professor Dr. Alfred Jante provided a test
field of assessment, theoretical or experi- technique which could be extended to all loop
mental, of scavenge flow and the references (or cross) scavenged two-stroke engines and
are reproduced here for completeness. (2) as such gave the two-stroke designer hope
(3) (4) (5) (6) (7) (8) (9) (10). This is far that at last there had emerged a method of
from an adequate reading list on the subject evaluating whether a particular power unit had
but it covers the basic background to the good or bad scavenging.
study of the flow of fresh charge into the The earlier paper by Blair (1) discusses
cylinder of a naturally-aspirated two-stroke the Jante method of scavenge flow assessment,
cycle engine; the type of engine commonly its advantages and drawbacks, and illustrates
used for motorcycles, chainsaws, outboard an experimental apparatus and test method which
motors and various other applications. In removes the tedium and inaccuracy associated
that previous paper the author (1) had pointed with the procedure given by Jante (9) .
out that while much of the literature provided Basically the Jante method involves motoring
valuable insight into scavenging as a fluid the engine with the cylinder head removed and
mechanic mechanism, little real design recording the velocity contour map of the
information accrued from a study of these often
excellent pieces of experimental work, real * Numbers in parentheses designate References
design information, that is, of applicability, at end of paper

In 1968 Professor Alfred Jante published

an SAE paper detailing a method of assessing terms of predicting good and bad scavenging
the scavenging behaviour of a two-cycle engine. for particular engine cylinders the results
It was a simple technique involving motoring have to be analysed rather more carefully and
the engine and measuring the (cylinder head completely than the approach adopted by Jante.
removed) velocity contours at the cylinder The experimental work describes motoring and
head level using pitot tubes. It attracted firing tests on one engine with many engine
wide attention in industry, but with varying cylinders of differing scavenging layouts and
degrees of acceptance and results. This correlates the scavenge tests motoring with
paper attempts to establish in a logical manner the firing tests which include measured scav-
and with a considerable volume of experimental enge and trapping efficiency values, cylinder
data that the method proposed by Jante has real pressure diagrams together with the usual
relevance, but to obtain acceptable accuracy in engine performance parameters.
231
0096-736X/81 /8901 -0231 $02.50
Copyright 1981 Society of Automotive Engineers, Inc.

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 232 G. P. BLAIR AND R. G. KENNY

flow at the top of the cylinder with pitot (g) each of the cylinders installed on the
tubes. Blair (1) shows how this data acquisi- engine for firing tests would have the trans-
tion and reduction may be accomplished using fer porting altered so that the scavenging
an automatic experimental sampling apparatus flow profiles were different, but that no
for the former and a digital computer and alteration be made so as to change the trans-
graph plotter for the latter. The identical fer port total area or timing or time-areas,
apparatus and the techniques for its employ- thus retaining identical gas dynamics cylin-
ment are used here in this study, so that a der to cylinder. In short neither motored
preliminary reading of that paper could be nor firing delivery ratios were to be effected
useful in aiding a fuller understanding of by transfer port modifications, only the
this report. directions with which the fresh charge entered
The papers by Jante (9) and Blair (1) the cylinder from the crankcase. Thence the
show many scavenge diagrams of this Jante firing tests conducted in (d) - (f) would
velocity contour type, but there is a lack of have identical gas dynamic characteristics
convincing evidence that 'good' diagrams do and the variation in scavenge behaviour would
produce 'good1 scavenging and vice-versa for be the "only" variable under scrutiny.
the 'bad1 scavenge pictures. Jante does (h) the same cylinders would be tested using
provide some experimental firing engine evi- the Jante method but with the test apparatus
dence but it is not complete; what is required described by Blair (1) which is by motoring
is an experimental study of a particular and the ensuing velocity contour maps and
engine where both firing performance character- derived data for symmetry, mixing, and strength
istics and motored Jante pictures can be of flow would be scrutinised for relevance to
compared unequivocally. This report hopes to the firing engine tests. Hopefully the Jante
fill in those gaps in our knowledge of this criteria of 'good' and 'bad' scavenging would
subj ect . be vindicated from this information, but at
least some useful information should emerge.
EXPERIMENTAL RATIONALE Even the information that the test method was
non-discriminatory would be positive in the
It would be pointless testing a whole sense that it would force researchers to look
series of engines for both firing performance for a superior research technique; it might
characteristics and motored Jante diagrams as be galling to admit such information but the
the dissimilar breathing in delivery ratio quest for knowledge has been known to progress
terms, and the unequal trapping behaviour down the occasional blind alley!
caused by varying unsteady gas dynamics in (i) the several cylinders would be tested for
the several exhaust systems, would mitigate scavenge diagrams, that is cylinder head level
against logical comparisons of fired and velocity contour maps under motoring condi-
motored engine scavenge data. It seemed to tions, and with chopped (pulsating) and with
the researchers at The Queen's University of steady air flow without piston motion so as
Belfast that what was required was the follow- to learn if there were correlations to be
ing: obtained between these various test methods
(a) one engine with many cylinders available and with the firing engine test results. At
(b) an engine of reasonable cylinder capac- the early prototype period it would be use-
ity so that instrumentation could be readily ful if the steady flow Jante method could be
attached used as a guide to the future transfer port
(c) the engine would be tested with each layout of the engine cylinder; it is a bit
cylinder in turn and with the same gas dyna- late in the day after the new cylinder has
mic breathing and trapping characteristics, been cast or fabricated to discover via either
that is with the same crankcase, carburettor, a motoring Jante test or a firing engine test
intake area and timing and exhaust system, that the engine has inferior scavenging. At
etc.
the beginning of design of a new cylinder
(d) the fired engine tests with each cylinder often little metal is in existence and the
would include scavenge and trapping efficiency execution of motoring Jante tests is as
together with torque, delivery ratio, air- impractical as firing engine tests whereas
fuel ratio, specific fuel consumption at mock-up models in clay or plastics could be
various engine speeds and fixed throttle evaluated using steady or chopped air flow,
settings Jante fashion; but is the test at all meaning-
(e) the fired engine tests with each cylinder ful under such conditions of "stationary"
as in (d) would also include further tests piston? If it is relevant is there an opti-
with different cylinder heads of varying geom- mum port opening or piston position?
etry but fixed compression ratio The entire test profile (a) - (i) has
(f) the fired engine tests with each cylinder been carried out over the last three years
as in (d) and (e) would also include the at The Queen's University of Belfast, but it
taking of cylinder pressure diagrams on consec- is not possible within the scope of a single
utive cycles for at least 500 cycles SAE paper to describe all of these experi-

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 SCAVENGING ANALYSIS 233

merits. A selection of tests is described here • Yamaha single-cylinder,

"torque induction",
as fully as is required for the reader to reach loop-scavenged, two-stroke

an adequate level of understanding of what has

taken place experimentally. Model
Bore
450
70.0 mm
Stroke 64.0 mm
Connecting rod centres 125.0 mm
EXPERIMENTAL PROCEDURE
Trapped compression ratio 6.9 : 1
Crankcase compression ratio 1.4 : 1
Exhaust port opens 88. 0o atdc
A single-cylinder engine, Yamaha type Transfer ports open 121.0° atdc
Exhaust port width 39.0 mm
DT250, used for the enduro motorcycle, was Radii on top corners of exhaust port 8.0 mm
selected for the test series and some twenty Radii on bottom corners of exhaust port 12.0 mm
Total effective transfer port width 89.0 mm
cylinders made available for modification Radii on top corners of transfer port 3.0 mm
Radii on bottom corners of transfer port 3.0 mm
purposes. A technical description of the Reed valve arrangement 60.0° wedge
engine appears in a table, Fig.l. It is a Number of reed petals 6
Length of reed petals 42.5 mm
spark-ignited, gasoline- burning, reed-induc- Clamped length of reed petals 5.0 mm
Length of reed ports 25.0 mm
tion valve, loop -scavenged, crankcase-compres- Width of reed petals 17.5 mm
sion, two-cycle power unit. The engine cylin- Width of reed petals 15.5 mm
Thickness of reed petals 0.25 mm
der is particularly suited for this experi- Reed material steel
Carburettor diameter 28.0 mm
mental task in that it has five transfer ports
which allow considerable variations to be
attempted for the alteration of the scavenge
flow. In this context, bearing in mind
experimental rationale (g) , with five trans- Fig. 1 - Technical data of the Yamaha DT250
fer ports present it is possible to vary engine
these port plans and elevation angles separat-
ely or together as well as the transfer duct
profiles so that many types of velocity
contour map may be obtained.
FIRING ENGINE TESTS - The DT250 engine
and transmission are assembled on a cradle,
see Fig. 2, and the drive to a jackshaft is
via a conventional final drive motorcycle
chain. The jackshaft is directly coupled to
a Schenck high-speed shaft thence to a
Schenck WS 130 eddy-current dynamometer with
constant speed controller and external torque
read-out. Thus all performance values of
torque and power are as recorded at the dyna-
mometer, and in motorcycle industry parlance
are "rear wheel" performance data. It is
estimated, although the authors cannot vali-
date this statement, that the crankshaft to
rear wheel performance loss due to friction
is some 7 - 10%. The engine is capable of
being controlled at ± 15 rev /min at any engine
speed by this measurement system. Fuel
consumption readings are recorded by an auto-
matic measurement system (designed and built
in-house) which allows for a recorded consump- Fig. 2 - The engine-transmission unit with
tion time of about 25 s for all fuel flow rates special cylinder head attached
with an accuracy of ± 1.0%. The air consump-
tion of the engine is recorded via a fixed also used here but the analytic devices have
orifice/box device designed to BS1042 (1966) been enhanced by the addition of a CO2 ana-
and has an accuracy of 0 to + 2%. The engine lyser, type Horiba Mexa 204 X,and this also
is tested at fixed throttle settings of full, simplifies the theoretical determination of
' and these values being based on area scavenge efficiency as given fully in Blair/
ratios of the full (1) 28 carburettor. Ashe (11); it is unnecessary to repeat that
The scavenging efficiency measurements analysis here. The experimental device is
are carried out using the apparatus originally an "overhead camshaft four-stroke valve" unit
described by Blair and Ashe (11) , but modi- geared at 11 : 1 from the crankshaft and thus
fied here to suit the DT250 engine. There is the engine operates for 10 normal cycles but
a very full description of the apparatus in upon the 'valve' opening on the 11th cycle,
that earlier paper, and it can be seen on the after trapping with ignition cut out, it is
cylinder head of this engine in Fig.l. The possible to expel some 20 - 40% of cylinder
recording devices described earlier (11) are contents for chemical examination. Similarly,

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 234 G. P. BLAIR AND R. G. KENNY

expansion gas samples can be obtained after procedure described by Blair (1) so there is
conventional combustion by re-timing the cam- no point in repeating that earlier paper. In
shaft appropriately but retaining ignition. addition the air-flow rates into the cylinder
The effectiveness of this device is two-fold. under motoring conditions were also monitored
Firstly the 'camshaft1 bolts to the cylinder so that delivery ratios recorded in the motor-
head, a specially machined head, but it allows ing mode could be related to those recorded
many cylinder heads and their profiles to be for the same cylinder under firing conditions.
tested without having to manufacture 'cam- This applied to tests at the several throttle
shafts' for each modified cylinder head. settings so that again correlation between
Secondly, the considerable fraction of cylinder motored and firing tests would be complete.
mass content expelled eliminates errors assoc- Equally, remembering rationale point (g) ,
iated with charge stratification which would should the experimental modifications produce
be incurred with solenoid type sampling different scavenge diagrams they should not
valves where typically 2 - 5% of trapped mass induce dissimilar delivery ratio curves in
is obtained for chemical analysis. (13; (14) either the motored or firing tests at the
The trapping efficiency values are deduced several throttle settings. By recording
from the measurements of oxygen levels in the delivery ratio values during motoring tests
exhaust gas. The deduction is made from the for the 'Jante pictures' this point could be
formulae postulated by Huber (12) where continuously checked during the progress of
trapping efficiency (TE) is given by the experimental work.

mp. = (21 - V02) * (21 - VU02 ( 1 - D)) STEADY AND UNSTEADY FLOW SCAVENGE
(21 - VU02) • (21 - V02(l - D)) TESTS - From the rationale point (i) tests for
scavenge diagrams were conducted with both
where V02 is the oxygen content in the dried steady air flow and with pulsating air flow,
exhaust gas (%) with the piston held stationary at several
VU02 is the oxygen content in the dried locations from transfer port opening to the
combustion products (%) bottom dead centre (or fully open) position.
D is a correction for the dried analy- Space limitations will prevent any discussion
sis. of this test series.
VU02 and D depend on the air/fuel ratio of
the test and Huber (12 - Fig. 7) gives a solu-
tion for a wide spectrum of air /fuel ratios.
Cylinder pressure diagrams are taken in EXPERIMENTAL RESULTS AND DISCUSSION

the cylinder head with a 12QP250 cv Vibro-

meter piezo-electric pressure transducer. Before embarking upon this section it
The signal of cylinder pressure, together will be useful to remind the reader that
with suitably accurate crankshaft position Jante (8) defined 'good' and 'bad' scavenge
markings, are recorded on a Philips Ana-Log 7 diagrams by the shape and position of the zero
tape recorder. The signals are converted on velocity line with respect to the plane of
an analogue to digital converter, UCC System symmetry of the scavenge flow. The basic
80, and processed in the University computer, definitions are presented in Fig. 3 where type
an ICL 1906S machine. From these rêcorded (a) shows the 'perfect' scavenge diagram. The
signals it is possible to deduce much statis- arrow at the top of the diagram shows the axis
tical information about combustion and the of symmetry of the scavenge flow and inher-
effect which variations in scavenging and ently, for most engines, the direction of
trapping have upon it. Further, as mentioned exhaust flow from the cylinder. The perfect
in rationale point (e) several cylinder heads scavenge velocity contour map has the zero
of constant trapped compression ratio are velocity line across the bore diameter with
tested. Space will prevent a discussion of lines of increasing velocity extending
the inter-relation between scavenging varia- smoothly towards the scavenge flow side with
tions and cylinder head profile but it should the highest velocity at the opposite side of
be mentioned that four cylinder heads were the cylinder from the exhaust port. The
tested; they were central spherical section negative 'velocity' on the exhaust half of the
chamber with and without a squish band, off- bore area shows the entrained flow and that
set chambers of spherical section with the caused by piston motion; inevitably any dimuni-
chamber offset over the exhaust port and also tion of that negative flow area by positive
over the scavenge flow side. Variations of flow pockets would indicate stray flow emanat-
spark plug positioning together with the ing from the fresh air or scavenge stream and
cylinder head changes were also included in would denote deteriorating scavenging. The
this section of the test series. 'perfect' scavenge picture, Fig. 3a, would have,
MOTORING ENGINE TESTS - The scavenge using the definitions given by Blair (1), the
diagrams obtained under motoring conditions following derived data from the pitot tube
are effected by using the same apparatus and mean velocity measurements:

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 SCAVENGING ANALYSIS 235

Flow Symmetry Ratio (FSR) = 1.0

Mixing Length Ratio (MLR) = 1.0
Mean Velocity (MV) = ? (depends on
delivery ratio?)

A further parameter has been used since

the earlier Blair (1) paper was written and it
is denoted as Scavenged Area Ratio (SAR) ,
defined as follows:

Z (area on scavenge half of cylinder

subtended by pitots recording positive
car = Fig. 3 - Fundamental Jante or scavenge
0.5 X BA patterns

Thus the 'perfect' scavenge diagram would have

a SAR value of unity.
The 'bad' diagrams of scavenging were
described by Jante as being unsymmetrical
(Fig. 3b), as having a 'tongue' pattern (Fig. 3c),
/y BDC
or forming a 'wall' pattern (Fig. 3d). Some
typical values for FSR, MLR and SAR are shown ' _ /50^
on Fig. 3 beneath each diagram, indicating how
these derived values may quantify the inferior
nature of the scavenging flow and how patterns
(b) , (c) and (d) have at least one derived
value straying significantly away from the
perfection value of unity. V
r'ìr' - 52 - -
At The Queen's University of Belfast over
the last ten years perhaps a 1,000 individual I
CYLINDER 8
pictures have been obtained from maybe 100
engines; the nearest to a 'perfect' picture
ever recorded is probably that in Blair Fig. 4a - Transfer port layout for cylinder 8
(1 - Fig. 11).
MOTORING TESTS - It is relatively obvious
that diagrams (a) (b) (c) (d) are different
and it requires little imagaination, helped by
the derived values of FSR, MLR, SAR and MV, to
arrive at a decision as to the superiority or ,
otherwise of such pictures. It is proposed y / ' BDC

now to show the motored and firing tests of

four cylinders which do not have such dis-
tinctly different scavenge diagrams (apparent-
ly) and yet have disparate firing performance
characteristics; these cylinders are designat-
ed by numbers being No. 8, No. 10, No. 12 and
No. 14. On Fig. 4 is shown the plan and main
transfer port elevation sections through these
several DT250 cylinders and the extent of the tV$> "
modifications carried out are clearly illus-
trated; they were accomplished using expoxy CYLINDE
fillers such as Devcon and re-machining. That
this was carried out so accurately as to Fig. 4b - Transfer port layout for cylinder 10
maintain the requisite constant delivery ratio
characteristics and scavenge symmetry (should
that also be desirable and that was the case On Figs. 6, 7, 8 and 9 are shown the
for these particular cylinders) is a tribute measured velocity contour maps for cylinders
to the craftmanship of the technicians at QUB 8, 10, 12 and 14 respectively at engine speeds
and their skill is here gratefully acknowledged. in each case of 3000, 4000, 5000 and 6000 rev/
On Fig. 5 is shown the motored delivery min and at full throttle. On Figs. 10, 11 and
ratio characteristics fc^r cylinders 10, 12 and 12 are shown the velocity contour maps for
14 at the several throttle openings and speeds cylinders 10, 12 and 14 respectively at one
and it can be seen that gas dynamic equality engine speed 4500 rev/min but at three throttle
has been achieved for these cylinders. positions, full, half and quarter throttle.

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 236 G- P- BLAIR AND R. G. KENNY

ah

l' Y
^
*°R J
/ ~1P)
1p^>r4rí^VN 38 R .1
' section ox s 60 R
section ox

CYLINDER 12 CYLINDER 14

Fig, 4c - Transfer port layout for cylinder 12 Fig. 4d - Transfer port layout for cylinder 14

On all scavenge diagrams the zero velocity line 0 9r

is drawn in heavily and the velocity contours
proceed on either side of it in equal incre- full throttle

ments of 5 m/s. The minimum and maximum veloc-

ities are stencilled on the computer/computer
graph plotter picture for ease of scrutiny and 0

discussion. The numeric data for these Figs. 07- CYL □ 12

6 - 12 is given in Table 1 for the derived CYL o 14

values of FSR, MV, MLR, SAR and DSSN.

The interesting point about these four
cylinders is that their scavenge diagrams are
apparently so similar. By Jante1 s definition Q ^
cylinder 8 looks quite good, perhaps the best,
yet cylinders 10 and 14 have higher SAR values >. ^
o) Motoring Tests
and cylinders 12 and 14 have superior MLR
a)
ratios; indeed cylinder 8 has the best symmetry 0-4 -
or FSR value. Another pointer could be that quarter throttle

the scavenge diagram for cylinder 8 alters

little with engine speed, but then neither
does cylinder 14 whereas cylinders 10 and 12
do not appear too good at 3000 rev/min. The RPM

partial-throttle pictures are also informa- 02 3000 3500 4000 4500 5000 5500 5000
tive in that cylinder 14 has a diagram which rev I mm

retains its profile at reduced delivery ratio,

a feature which cannot be claimed for cylinder
12 where the quarter-throttle picture has a Fig. 5 - Motoring delivery ratios for cylinders
decided 'wall pattern' look to it. The debate 10, 12, 14
on inferiority or superiority of these four
cylinders, one to the other, could proceed at
length but the arbiter is available for a delivery ratios at which the engine was motored
decision, namely the firing engine tests. for scavenge diagram evaluation were quite
FIRING TESTS - On Fig. 13 is shown the representative of those found during real
delivery ratio characteristics for all four firing conditions at equal conditions of speed
cylinders 8, 10, 12 and 14 with respect to and throttle position. Equally (remember
speed and throttle opening. It is particu- rationale point (g)) the modifications to
larly noticeable that the only curve to be cylinders 8, 10, 12 and 14 did not interfere
considerably modified by unsteady gas dynamics with delivery ratio characteristics; that the
phenomena, from the exhaust pipe particularly, scavenge behaviour is different is clear (or
is that for full throttle. The exhaust pipe not so clear) from Figs! 6 - 12.
on this engine, although containing an effec- On Figs. 14, 15 and 16 are shown the
tive silencer for street use, is also of the brake mean effective pressure, brake specific
tuned type. Nevertheless it is clear that the fuel consumption and air-fuel ratio perform-

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
SCAVENGING ANALYSIS 237

Fig. 6 - Scavenging pictures at full throttle

for cylinder 8

Fig. 7 - Scavenging pictures at full throttle

for cylinder 10

Fig. 8 - Scavenging pictures at full throttle

for cylinder 12

Fig. 9 - Scavenging pictures at full throttle

for cylinder 14

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 238 G. P. BLAIR AND R. G. KENNY

Fig. 10 - Scavenging pictures at several Fig. 11 - Scavenging pictures at several

throttle openings, cylinder 10 throttle openings, cylinder 12

0 9
CYL • 8
CYL * 10
CYL □ 12 fu" /throttle
°8

Fig. 12 - Scavenging pictures at several

throttle openings, cylinder 14 quarter throttle

0 4 -

ance parameters with respect to engine speed 0 3 L- j

and for full throttle, half throttle and 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000
RPM engine speed rev /min
quarter throttle respectively; all four cylin-
ders are drawn together on the same diagrams.
It is now very clear that the best cylinder is Fig. 13 - Firing delivery ratios for cylinders
No. 14, followed by No. 8, No. 10 and No. 12. 8, 10, 12,14
This is almost completely true whether it be
torque (or bmep) or fuel consumption (as bsfc)
as the criterion of excellence, or whatever
the throttle setting. It will be noted that
the experimenters have controlled air-fuel Table 1 - Numeric Data for Figs. 6 - 12

ratio variation from cylinder to cylinder at FIG CYL TAR RPM FSR MLR SAR MV DSSN

equal speed points within acceptably fine

681 3000 0.93 1.20 0.58 11.7 2.45
limits for a jet-type slide-valve carburettor 681 4000 0.84 1.14 0.60 15.5 2.26
set-up. It will also be observed that while 681 5000 0.93 1.11 0.66 18.0 2.11
6 8 1 6000 0.91 1.12 0.64 18.2 1.81
cylinder No. 14 is the best for bmep and bsfc
at virtually all speed points at full throttle, 7
7
10
10
1
1
3000
4000
0.84
0.72
1.42
1.29
0.67
0.67
10.0
12.4
2.09
1.81
the advantage is more clearly observed in the 7 10 1 5000 0.54 1.14 0.66 14.6 1.71

speed band where the tuned exhaust pipe has 7 10 1 6000 0.67 1.45 0.65 15.6 1.55

its greatest effect, that is the 1000 rev /min 8 12 1 3000 0.72 1.17 0.59 9.7 2.03
8 12 1 4000 0.45 1.72 0.70 13.6 1.98
either side of peak torque. There is a 15% 8 12 1 5000 0.73 1.08 0.63 14.1 1.65

advantage for cylinder 14 over cylinder 12 for 8 12 1 6000 0.65 1.06 0.57 13.2 1.32

bmep and some 20% superiority in specific fuel 9 14 1 3000 0.64 1.21 0.64 11.9 2.49
9 14 1 4000 0.51 1.26 0.65 15.4 2.25
consumption for the same pair of cylinders, at 9 14 1 5000 0.64 1.05 0.65 19.0 2.23
the full-throttle test condition at 6000 rev/ 9 14 1 6000 0.78 1.10 0.68 20.5 2.42

min. The percentage superiority would be 10 10 1 4500 0.68 1.15 0.72 13.7 1.74

greater at 7000 rev/min, and somewhat less at 10 10 | 4500 0.62 1.16 0.57 8.5 1.66
10 10 I 4500 0.52 1.13 0.44 5.0 1.58
peak torque at 5500 rev/min. At half-throttle
11 12 1 4500 0.63 1.08 0.61 12.7 1.65
these trends are still evident. It is at
11 12 i 4500 0.65 1.17 0.53 7.6 1.51
quarter- throttle that one observes some fascin- 11 12 I 4500 0.89 1.30 0.37 4.3 1.38

ating results; consider the situation at 12 14 1 4500 0.65 1.19 0.64 16.4 2.18

4500 rev/min where the torque advantage of 12 14 i 4500 0.65

12 14 I 4500 0.62 1.05 0.43 6.7 2.12
1.17 0.60 11.3 2.21

cylinder 14 over cylinder 12 is very marked at

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
SCAVENGING ANALYSIS 239

600"
500-

A
¿50-
O jP m

sT o

•-_/-^ sT 30°l* o
30°l*
N' ļ°65
Y*

^^'SŠL half throttle

N?' 060

>ss

^ N
1»^ O

o 15r 0Z>5^ O ^
° 12i.

2000 2500 3000 3500 ¿000 ¿500 5000 5500 6000 65C0 7000

! 13 - ^ Joxo engine speed rev /min

B 12I Fig. 15 - Torque and fuel consumption curves

250 0 3000 3500 ¿000 ¿500 5000 5500 6000 6500 7000
at half throttle
engine speed rev /min

Fig. 14 - Torque and fuel consumption curves

at full throttle

¿00 -

70 .

350 -

300 -
CL bmep

I 200 r^NS^2^
200 r

0 70 50 ^ °2 ™
ISO*- . ..
quarter . throttle ..

/x^^' "065
nN ' ? / a'^ ' 10 F 0 50 ^
$qD dUQI"ter thottle

Vf^lj nN ' ? / a'^ ' 10 F 0 50 ^

p - a A/F
jT L 70 " /llv " TRAPPING EFFICIENCY CURVES

^ RPM r^^Cļj 60P

+y rpm

0 2000 2500 3000 3500 ¿000 ¿500 5000 5500 6000 6500 2500 3000 3500 ¿000 ¿500 5000 5500 6000 6500 7000
engine speed rev /min engine speed rev / min

Fig. 16 - Torque and fuel consumption curves Fig. 17 - Trapping efficiency curves at various
at quarter throttle throttle openings

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 240 G. P. BLAIR AND R. G. KENNY

39% while the fuel consumption differential for 090-

these two cylinders is some 25%. Successful

part-throttle operation is very heavily influ- 0Ö5"

enced by scavenging and this is clearly shown

by these test results. Could the reader be
reminded that these tests are being conducted
at equal conditions of delivery ratio otherwise
these comparisons would be greatly reduced in 0 75 - 60OÒ~RPNT' *
validity?; it is rationale (g) again.
DR °
On Fig. 17 is shown the trapping efficiency 070 095 Õ4Õ Õ45 Õ5Õ 055 tfW Uts Oft)
TE results for the four cylinders 8, 10, 12
delivery ratio
and 14 for the three fixed throttle settings
over the speed range 2500 - 7000 rev/min. The
clear superiority of cylinders 14 and 8 over Fig. 18 - Variation of trapping efficiency with
cylinders 10 and 12 is obvious; some 8% of delivery ratio
actual trapping efficiency differential or
about 10% of real improvement is evident at
6000 - 6500 rev/min at full-throttle.
On Fig. 18 is illustrated the relationship 0 875 r
of trapping efficiency TE with delivery ratio
^ scavenging efficiency
DR at 6000 rev/min for the four cylinders;
this figure serves to underscore the comments 0850-
'

passed above.
The results of the scavenging efficiency
SE measurements are given on Fig. 19 where the
^SSSSsD^^' 0 ~~~ 0
trends already deduced are repeated. The 0825 " fuit throttle ^
figure is taken from data recorded at full-
throttle for the four cylinders. The surpris-
RPM °

OQOCm

2500 3000 3500 4000 4500 5000 5500 6000 6500 7000
ing element about this data is no longer that
engine speed rev / min
cylinder superiority is, in descending order,
No. 14, No. 8, No. 10 and lastly No. 12, but that
the differential between them for the absolute Fig. 19 - Scavenging efficiency curves at full
value of scavenging efficiency is so small; throttle
but two percentage points separate the best
from the worst. Further, this differential
while small extends virtually as a constant £qqq exhaust spark tdc exhaust
close ; ; opęn
throughout the speed range. This is in
3600 - I I I ļ
contrast to almost all of the other perform-
ance parameters, including trapping effici- 3200 (- ! ! ! I
ency where the advantage (or disadvantage) 2800 - ! I ! M ¡
differential increases as some function of
speed (or is it total air flow?). Yet another 0 uoo- ¡ ; !/ Vo 12 ;
interesting facet of the results is that the
value of ^scavenging efficiency is so constant î 2000 ' I i '
throughout the- speed range at full-throttle, a. 1600 - i ! /! !
for example the value for cylinder 14 falls
just three percentage points from 2500 - 7000 1 120°- ! !/ i ' '
rev/min. 800 - ¡ / ! ļ
The cylinder pressure diagrams obtained
¿oo - ! jr ' '
in this study could form a complete SAE paper
by themselves so it will suffice to show but °0 3 0 50 9 0 120 1 50 180 210 2¿0 270 300 330 360
one example to reinforce some of the points degrees after bottom dead centre

made earlier with regard to quarter-throttle

performance. On Fig. 16 it was observed that Fig. 20 - Cylinder pressure diagrams at ¿
there was a considerable disparity at 4500 throttle at 4500 rev/min
rev/min between the performances of cylinders
10 and 12 by comparison with the best cylinder, tests in this report the cylinder head used
No. 14. The cylinder pressure diagrams for was central chamber with a squish band and
these same cylinders at that speed are plotted is as sketched in Fig. 21. It is clear from
on Fig. 20, and the values plotted are those Fig. 20 from whence comes the better bmep/imep
averaged arithmetically over 500 consecutive performance of cylinder 14 over the other two
cycles; the cylinder head employed in this cylinders for the rate of pressure rise and
case being a central chamber with the spark peak pressure is clearly superior for the
plug located over the exhaust side. For all former cylinder.

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 SCAVENGING ANALYSIS 241

580r

spark plug sample valve ©

location ' / location 570-

' ! I 560 - ♦17

550 L (g)*16
ì 54
g- +11 full throttle
E 6000 rev I min

530- +9
©7
520-

<t> 54 1 510-

exhaust side scavenge side @

500l

13 U 15 16 17 1ft 19 20 21
^
mean velocity MV m/s

cross section in plane of symmetry

Fig. 22a - Torque related to mean velocity MV,

Fig. 21 - Cylinder head used for firing tests 6000 rev/min
reported

CORRELATION OF SCAVENGE DIAGRAMS AND

PERFORMANCE CHARACTERISTICS - The four
cylinders were chosen for presentation in this 570-

report quite deliberately so that the reader

could be confronted with the same dilemma 560- -17

faced by the researchers. Remember that the

550- (C)*16
scavenge diagrams Figs. 6 - 12 apparently a?
showed no absolutely clear distinction between 540- -13
g. +11
cylinders 8, 10, 12 and 14; certainly cylin- E
n

der 14 had no obvious claim to the premier ST) - full throttle

♦ 9 6000 rev/ min
position. The performance measurements
Figs. 13 - 21 were quite unanimous in declaring
520-

cylinder 14 the "winner". Thus if the Jante 510-

method was to have any applicability for the ©

assessment of scavenge behaviour then some 5001

130 1 AO 1 50 1 60 1 70 1 80 1 90 2 00 2 10
information must be available from those DSSN

scavenge diagrams or their derived values.

Should that information not be forthcoming
then the Jante method would have to be declared Fig. 22b - Torque related to DSSN, 6000 rev/min
to be no longer a useful development approach
for the industrial environment. are circled on Fig. 22(a) for ease of recogni-
It has already been pointed out that there tion. A supporting view is obtained by plott-
was no really conclusive information on the ing on Fig. 23 (a) the identical information but
cylinder order of merit from the shape of the with respect to fuel consumption (as bsfc) .
pictures themselves, nor from the derived The correlation between mean velocity MV and
values for symmetry FSR, mixing MLR or area the performance parameters of torque and fuel
scavenged SAR. Attention was paid to that consumption at this speed, which is close to
other criterion, mean velocity MV, a value peak power, is very good indeed not only for
which up to that time had been paid scant atten- the four selected cylinders but for another
tion at QUB but which was always deduced on seven cylinders as well.
the computer output more as a matter of course The situation when the mean velocity MV
than a matter of necessity. As in Blair (1) is plotted against bmep at 5000, 4000 and
the mean velocity is defined as 3000 rev/min at full throttle for the four
cylinders, and others, can be observed in
I(PPV X PSA) , Figs. 24(a), 25(a) and 26(a) respectively. As
™ - 0.5 X X BA m/s , the performance characteristics bunch together
below about 4500 rev/min it would be expected
On Fig. 22(a) is drawn the relationship between that the correlation would not be so effective.
mean velocity MV and torque (as bmep) at full However, it does just that, in fact it would
throttle at 6000 rev/min for not only the four appear to be a quite sensitive indicator of
cylinders 8, 10, 12 and 14 but also for several potential scavenging performance. Accepting
other cylinders. The relevant four cylinders that thesis, there has to be some relationship

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 242 G- P. BLAIR AND R. G. KENNY

0X9 r
049

© ©
048- full throttle
048 - full throttle
6000 rev / min
6Q0Q rev / min

0X7-
0 47 -
♦7

@ 0X6-
©•'
£ 0 X6 -
I £
♦9 * +9
S' 0X5- +11 o»
jí045
♦"- ♦"
iiii
O *16

"ou- .,3 £ OU - 13
© ©
043-
0 X3 -

♦17 ♦17

0X2- 042-

.15 © .15 ©
0-4,1 13 14 15 16 17 18 19 20 21 °'113 It 15 16 17 18 19 2 0 21
mean velocity MV m/s DSSN

Fig. 23a - Fuel consumption related to mean Fig. 23b - Fuel consumption related to DSSN,
velocity MV, 6000 rev /min 6000 rev /min

580 ■
580r

full throttle

570 - full throttle 570 ■ 5000 rev /mm

5000 rev /min

560-
560-

550- 550-
o
CL
0
Q.
-* 540- O.540 -

1 53o .17 .15 © S53C. .15 ©

© ©
523- *7 g 52°" ^ ł?-9
© ©^
5Ï- 510-

50C
•11 © im *fl .©
#5 iSī SÍ!! 5vü 5Í¡5 30
140 145 150 155 160 16 5 170 175 18 0 18 5 190
DSSN
mean velocity MV m/s

Fig. 24a - Torque related to mean velocity MV,

5000 rev /min Fig. 24b - Torque related to DSSN, 5000 rev/min

480
480 r
full throttle
470 - 4000 rev/min full throttle

470 - 4000 rev/min

460 -
460 -
^ »15
«0- © ^ ^ *15
* .7 © 450- ^ @
a 440 - 4i7 ♦ 7 @ W
& U0- .17 ♦ 7 W
1 430 ^ +13 16
430 ^ JO) ł9
JSC

I 430 @ ^ >13 J6
420 -

420-
©W
.410- W
410 - ^ ©J1
^

40°124 12 8 Ī32 Ī36 Ī40 ŪX 148 15 2 156

mean velocity MV m/s xm I

180 19 0 20 0 210 22 0 2 3 0
DSSN

Fig. 25a - Torque related to mean velocity MV,

4000 rev/min Fig. 25b - Torque related to DSSN, 4000 rev/min

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 SCAVENGING ANALYSIS 243

460r
460 •
full throttle
3000 rev /min
450-
450 - full throttle
3000 rev/min
♦15
♦15
440-
440 -

©W
430 - ©A W
S 430 - © ® ©A
420- o 420-
© ®
! © w * (ft) ^ ł9
410- w „13
1 410 - ^ »13
400 ■
♦11
400 - *11

390- 390-

3 96 Ī00 Ī04 Ī08 ĪT2 ītē Ī20 Ī2 4 3 20 ÏM 22 2^3 24 2^5 26

mean velocity MV m/s D5SN

Fig. 26a - Torque related to mean velocity MV, Fig. 26b - Torque related to DSSN, 3000 rev/min
3000 rev/min

27- 23(b), 24(b), 25(b) and 26(b) is plotted this

26- 'dimensionless scavenge strength number1 DSSN
25-
and it can be seen that as the delivery ratios
24-
DR are virtually constant at full throttle for
2 3 v/ the various tests there is little variation in
the pattern produced by comparison with that
z
21 -
created by the mean velocity MV numbers.
g20" ¿»WO RPM Nevertheless the trend illustrated by DSSN is
19-
evident in the performance characteristics.
18 -
The data from Figs. 10 - 12 and other data
not listed are used to compile values of DSSN
against delivery ratio DR at the fixed speed of
4500 rev/min; this is shown in Fig. 27 for
cylinders 10, 12 and 14 and, apart from the
1 Q3 04 05 05 07 06
delivery ratio DR obvious superiority of cylinder 14 and the
inferiority of cylinder 12, it is clear that
Fig. 27 - Scavenge strength related to air-flow the dimensionless scavenge strength number
rate DSSN is virtually a constant with delivery
ratio DR for any particular scavenge layout.
Further, the lower the delivery ratio the
between the flow on the, positive (scavenge) better is the cylinder-to-cylinder separation
side of the picture and the air mass flow rate seen from the DSSN value. On the other hand
pumped through the engine crankcase and into when the full- throttle data from Figs. 6 - 9 is
the cylinder. The objective would be to analysed to produce DSSN values and plotted
create a dimensionless number representative against engine speed RPM, the slower response
of "scavenge strength", which the mean time of the pitot tubes at higher speeds is
velocity MV provides dimensionally , so that all reflected in a number change with engine
cylinders on all engines at all engine speeds speed. This data is plotted on Fig. 28, DSSN
and throttle openings could be related in some versus RPM for cylinders 8, 10, 12 and 14, and
manner. While this may be ambitious in general, while the order of merit is evident at 5000
at least for a particular cylinder it could and 6000 rev/min for the several cylinders it
have relevance for comparison of cylinder modi- is not so clear below 4500 rev/min. However
fications which involve delivery ratio DR cylinders 14 and 8 are demonstrably superior
changes at fixed speed and throttle opening. to cylinders 10 and 12 at any speed but the
The authors would like to propose the following differentiation between the former and latter
"dimensionless scavenge strength number" DSSN pair of cylinders is only clear above 4500
rev/min.
w^ere
DSSN
Taking all of these facts together,
30 X MV particularly those illustrated on Figs. 22(b),
DR X STRK X RPM 23(b), 24(b), 25(b), 26(b), 27 and 28. The
following comments can be made
The value of delivery ratio DR would be that (i) at any given speed the DSSN value will
value measured during the period of the Jante cater for engine development modifications
test at the fixed speed RPM. On Figs. 22(b), which will produce delivery ratio DR changes

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 244 G. P. BLAIR AND R. G. KENNY

as well as scavenge flow alterations,

27r
(ii) if a speed for engine development has 26 -
to be chosen for motoring Jante tests it would
2 5
appear that the best speed to use would be
either peak torque speed or peak power speed
or both,
(iii) that motoring scavenge testing at
reduced throttle openings and delivery ratios
is an excellent guide to the resultant perform-
ance characteristics. I20
19 X.

CONCLUSIONS

Since Jante (9) published his original

paper there has been a continuous technical
discussion as to the effectiveness of this ^'l2
approach for the improvement of scavenging of engine speed RPM
the two-stroke engine. Hopefully this report 2000 3000 4000 5000 5000

will clarify some hitherto unsupported postula-

tions, among them being the necessity or
otherwise of recording the data completely and
analysing it appropriately. It is clear that Fig. 28 - Scavenge strength related to engine
rudimentary sketching of the Jante diagrams speed
for cylinders 8, 10, 12 and 14 would have
yielded little useful information and would
probably have convinced the experimenter, when the performance characteristics for the four
the performance characteristics were obtained, cylinders described, that good scavenging is
that the Jante method failed to differentiate essential for optimum performance to be
between łgoodT and fbadł scavenging. It is attained by a two-stroke cycle engine. Properly
equally clear that when the data from the carried out, the Jante method seems to be the
Jante method is analysed correctly it is cap- most effective experimental approach suggested
able of providing just that differentiation thus far for the improvement of scavenging of
which the researcher and the engine developer a two-cycle power unit.
requires. Therefore the main conclusion which
can be drawn is that the data from the measur- NOMENCLATURE
ing apparatus should be acquired as automatic-
ally as possible and analysed in the same PPV - positive pitot velocity m/s
manner; with the advent of cheap personal NPV - negative pitot velocity m/s
computers and programmable calculators there PSA - area subtended by the pitot m2
is no logical barrier for the engineer in DR - delivery ratio )
executing such procedures. It would appear TE - trapping efficiency )see Blair/Ashe (11)
that the derivation of such numeric data as SE - scavenging efficiency)
flow symmetry ratio FSR, mixing length ratio bmep - brake mean effective pressure kPa
MLR, scavenged area ratio SAR and, most import- bsfc - brake specific fuel consumption
ant of all, the dimensionless scavenge strength kg /kWh
number DSSN are as essential as the drawing FSR - flow symmetry ratio)
of the velocity contour map itself; the latter MLR - mixing length ratio) see Blair (1)
criterion DSSN demands that the delivery ratio MV - mean velocity m/s )
DR be measured together with engine speed RPM SAR - scavenged area ratio
for the production of that data value. The DSSN - dimensionless scavenge strength number
requirement that these extra values be obtained RPM - engine speed rev/min
for a Jante method scavenging evaluation may SV - cylinder swept volume m3
seem to many to transfer an otherwise simple BA - cylinder bore area m2
process into an experimental and arithmetic STRK - cylinder stroke m
nightmare but, as already mentioned, with modern A/F - air to fuel ratio
computer and personal computer technology it TAR - throttle area ratio
is neither complex nor expensive. CYL - cylinder nomenclature number
The report shows the range of performance
characteristic variation which is possible with ACKNOWLEDGEMENTS

fgoodT and 'bad* scavenging, indeed it is

doubtful if the researchers actually achieved The authors acknowledge their debt to SRC
truly bad scavenging, so the differential could and Queen1 s University of Belfast for providing
be even greater than that illustrated here. the experimental and computational facilities
There cannot be much doubt, when one examines without which this work could not be carried

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms
 SCAVENGING ANALYSIS 245

out. The industrial sponsors of this work must 6. N. Dedeoglu, "Model Investigations on
be thanked for their generous financial and Scavenging and Mixture Formation in the Dual-
technical support; they are Yamaha Japan, Fuel or Gas Engine". Sulzer Tech. Rev, Vol.51,
Mercury USA, Bombardier-Rotax Austria, AMF USA, No. 3 (1969), p. 133.
Kohler USA, Motobecane France, Ossa Spain, 7. A. Kannapan, "On Model Investigations
Polaris USA, Roper USA and Piaggio Italy. pertaining to Scavenging in Two-Stroke Diesel
Yamaha Motor Japan deserve special thanks Engines". ASME Paper 71-DGP-8, April 18-22,
for their liberal supply of engines, spare 1971.

cylinders and spare parts for the execution of

the experimental work. 8. H. Kothe, "Effect of Non-Symmetrical
The Head of Department, Professor Bernard Scavenging on the Performance of Crankcase-
Cross land, F.Eng., F.R.S., needs special men- Scavenged Two-Stroke Cycle Spark Ignition
tion for the way in which he encourages such Engines", Kraf tf ahrzeugtechnik, Vol.1, (1985),
experimental research work. p. 3. (See also Vol.4 (1960), p. 123).
9. A. Jante, Scavenging Flow Ducts of
REFERENCES
3-Cyl 2-stroke Cycle Engines". SAE Transactions,
Vol.77 (1968), paper No. 680468.
1. G. P. Blair, "Studying Scavenge Flow
in a Two-Stroke Cycle Engine". SAE paper
10. T. Aoyama, et al, "Abnormal Combustion
No. 750752, 1975.
2. W. Rizk, "Experimental Studies of the in Two-Cycle Motorcycle Engines", Bull. J.S.A.E.,
November 1974, No. 57, p. 433.
Mixing Processes and Flow Configurations in
11. G. P. Blair and M. C. Ashe, The
Two-Cycle Engine Scavenging". Proc. I. Mech. E.,
Unsteady Gas Exchange Characteristics of a
Vol.172, (1958), p. 417.
Two-Cycle Engine", SAE Paper No. 760644, 1976.
3. H. Sammons, A Single-Cycle Test
12. E. W. Huber, Measuring the Trapping
Apparatus for Studying 'Loop Scavenging' in a
Efficiency of Internal Combustion Engines
Two-Stroke Engine". Proc. I. Mech. E., Vol.160,
Through Continuous Exhaust Gas Analysis".
p. 233.
SAE paper No. 710144, 1971.
4. W. H. Percival, Method of Scavenging
Analysis for Two-Stroke-Cycle Diesel Cylinders". 13. T. Asunuma and S. Yanagihara, "Gas
SAE Cleveland Meeting, October 26, 1954. Sampling Valve for Measuring Scavenging Effic-
5. C. Martini and M. Oggero, Experimental iency in High-Speed Two-Stroke Engines", SAE
Methods for the Discovery of Two-Stroke Engine Transactions, Vol.70, p. 420, paper T-47, 1962.
Scavenging". Paper 710145, presented at SAE 14. R. R. Booy, Evaluating Scavenging
Automotive Engineering Congress, Detroit, Efficiency of Two-Cycle Gasoline Engines",
January 1971. SAE paper No. 670029, 1967.

This content downloaded from 91.228.42.246 on Sat, 02 Oct 2021 05:27:18 UTC
All use subject to https://about.jstor.org/terms