Download the full version of the ebook at

https://ebookultra.com

The Intelligent Movement Machine An

Ethological Perspective on the Primate Motor
System 1st Edition Michael Graziano

https://ebookultra.com/download/the-intelligent-
movement-machine-an-ethological-perspective-on-
the-primate-motor-system-1st-edition-michael-
graziano/

Explore and download more ebook at https://ebookultra.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Lectures on Perception An Ecological Perspective Michael T

Turvey

https://ebookultra.com/download/lectures-on-perception-an-ecological-
perspective-michael-t-turvey/

ebookultra.com

Consciousness and the Social Brain 1st Edition Michael S.

A. Graziano

https://ebookultra.com/download/consciousness-and-the-social-
brain-1st-edition-michael-s-a-graziano/

ebookultra.com

Machine Learning An Algorithmic Perspective Second Edition

Stephen Marsland

https://ebookultra.com/download/machine-learning-an-algorithmic-
perspective-second-edition-stephen-marsland/

ebookultra.com

Intelligent Adaptive Systems An Interaction Centered

Design Perspective 1st Edition Ming Hou

https://ebookultra.com/download/intelligent-adaptive-systems-an-
interaction-centered-design-perspective-1st-edition-ming-hou/

ebookultra.com
Intelligent Adaptive Systems An Interaction Centered
Design Perspective 1st Edition Ming Hou (Author)

https://ebookultra.com/download/intelligent-adaptive-systems-an-
interaction-centered-design-perspective-1st-edition-ming-hou-author/

ebookultra.com

The Neurocognition of Dance Mind Movement and Motor Skills

1st Edition Bettina Bläsing

https://ebookultra.com/download/the-neurocognition-of-dance-mind-
movement-and-motor-skills-1st-edition-bettina-blasing/

ebookultra.com

White Collar Crime An Opportunity Perspective 1st Edition

Michael Benson

https://ebookultra.com/download/white-collar-crime-an-opportunity-
perspective-1st-edition-michael-benson/

ebookultra.com

Biomechanics and Motor Control of Human Movement 4th

Edition David A. Winter

https://ebookultra.com/download/biomechanics-and-motor-control-of-
human-movement-4th-edition-david-a-winter/

ebookultra.com

Primate Anatomy Third Edition An Introduction Friderun

Ankel-Simons

https://ebookultra.com/download/primate-anatomy-third-edition-an-
introduction-friderun-ankel-simons/

ebookultra.com
The Intelligent Movement Machine An Ethological
Perspective on the Primate Motor System 1st Edition
Michael Graziano Digital Instant Download
Author(s): Michael Graziano
ISBN(s): 9780199716791, 019971679X
Edition: 1
File Details: PDF, 11.56 MB
Year: 2008
Language: english
The Intelligent Movement Machine
This page intentionally left blank
The Intelligent
Movement Machine
An Ethological Perspective
on the Primate Motor System

MICHAEL S. A. GRAZIANO, Ph.D.

1
2009
1
Oxford University Press, Inc., publishes works that further
Oxford University’s objective of excellence
in research, scholarship, and education.

Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2009 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016
www.oup.com

Oxford is a registered trademark of Oxford University Press

All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Graziano, Michael S. A., 1967–
The intelligent movement machine : an ethological perspective on the primate motor system /
Michael S.A. Graziano.
p. ; cm.
Includes bibliographical references and index.
ISBN-13: 978-0-19-532670-3
1. Motor cortex—Physiology. 2. Human locomotion. 3. Animal locomotion.
4. Primates—Physiology. I. Title.
[DNLM: 1. Motor Cortex—physiology. 2. Movement—physiology. 3. Models, Animal.
4. Primates—physiology. WL 307 G785i2009]
QP383.G78 2009
612.8'252—dc22
2008012835

9 8 7 6 5 4 3 2 1
Printed in the United States of America
on acid-free paper
Acknowledgments

This book would not have been possible without many collaborators and col-
leagues including Charlie Gross, my mentor through many years; Charlotte
Taylor and Tirin Moore, who collaborated on the initial experiments in which
we electrically stimulated the monkey motor cortex on a behaviorally relevant
time scale; Dylan Cooke who studied the motor cortex control of defensive
movements; Tyson Aflalo who explored the computational and theoretical
implications of the data; Sabine Kastner and Jeffrey Meier who made it possi-
ble for my lab to explore functional imaging in the human motor cortex; Tyler
Clark who acquired many books and articles that went into the research; and
Theodore Mole who remains an unfailing presence in the lab. In addition, Ed
Tehovnik, Sabine Kastner, Charlie Gross, and Craig Panner made many useful
comments on the manuscript.

v
This page intentionally left blank
Preface

In 1870, Fritsch and Hitzig discovered the motor cortex in the dog brain. Since
then, for one hundred and thirty years, researchers have grappled with the
fundamental question of motor cortex: How is it organized?
I believe this question is finally answered. The answer is simple in concept.
An animal’s normal movement repertoire is flattened onto the cortical sur-
face. The complexity of the map comes from the complexity of the movement
repertoire. With a good description of the typical movement repertoire of a
species of animal, it should be possible to predict mathematically the layout of
the motor cortex. We now have an approximate description of the movement
repertoire of macaque monkeys, and with it we can explain the overarching
organization of the monkey motor cortex.
The theory that the motor repertoire is flattened onto the motor cortex is
one specific example of a general principle of brain organization. One might
say that the mental repertoire of the animal is mapped somehow onto the
entire brain. In the case of movement, the repertoire is conveniently observ-
able and therefore its mapping onto the cortical surface can be studied directly.
The purpose of this book is to review experiments on how the motor reper-
toire is mapped onto the cortex, ranging from the initial discovery of motor
cortex to the present.

vii
This page intentionally left blank
Contents

CHAPTER 1: Introduction 3

CHAPTER 2: Early experiments on motor cortex 13

CHAPTER 3: An integrative map of the body 39

CHAPTER 4: Hierarchy in the cortical motor system 51

CHAPTER 5: Neuronal control of movement 71

CHAPTER 6: What can be learned from electrical stimulation? 85

CHAPTER 7: Complex movements evoked by electrical stimulation of

motor cortex 97

CHAPTER 8: The match between natural neuronal properties and

stimulation-evoked movement 125

CHAPTER 9: The movement repertoire of monkeys 139

CHAPTER 10: Dimensionality reduction as a theory of motor cortex

organization 151

CHAPTER 11: Feedback remapping and the cortical-spinal-muscular

system 167

CHAPTER 12: Social implications of motor control 181

Literature Cited 199

Index 219

ix
This page intentionally left blank
The Intelligent Movement Machine
This page intentionally left blank
Chapter 1

Introduction

BACK STORY: MIXING TWO EXPERIMENTAL CULTURES

When I was a postdoc at Princeton University, I worked on the integration of
vision, touch, and movement in the monkey brain. The experiments involved
monitoring the activity of single neurons in the motor cortex during the mon-
key’s movements or during the presentation of sensory stimuli. I had never
considered using electrical stimulation to study motor cortex. Many colleagues
had suggested the technique, and my response was something like, “You get
muscle twitches. Big deal. You can’t really learn anything.”
At the same time another postdoc in the lab, Tirin Moore, had begun a set
of experiments on the frontal eye field of monkeys. He used a common elec-
trical stimulation technique in which pulses of current are delivered into the
cortex through a fine, hair-like electrode. The pulses are presented in a train at
high frequency (typically 200 pulses per second). This method directly activates
a small sphere of brain tissue around the electrode tip. The directly stimulated
neurons then recruit physiologically connected networks. If stimulation is
applied to a spot in the frontal eye field, it evokes an eye movement that closely
resembles a natural one.
When experimenting on the frontal eye field, for each new monkey stud-
ied, one typically first explores a broad area of cortex, stimulating in a variety
of locations to find the borders of the area of interest. During one such explo-
ration, Tirin came running down the hall to my office, his lab coat billowing
behind him like the cape of a superhero, and said, “Mike, you have to look at
this.” I came and looked.
He held a button in his hand, and every time he pressed the button, the
monkey sitting in the plastic monkey chair in the center of the room extended
his arm forward and shaped his fingers as if reaching for something invisible.
The effect was immediate, consistent, and obviously as amazing to the monkey
as it was to us because the monkey grabbed hold of his hand with the other
one, pulled it straight down, and sat on it, effectively ending the experiment
for the day. Tirin had obviously missed the frontal eye field and gotten the
electrode into the primary or premotor cortex.
“We have got to study this,” he said.
The evoked movement was no muscle twitch. The reason was immediately
obvious to us. In a standard stimulation experiment on motor cortex, the
stimulation is applied in a brief burst for 50 ms or less. The result of this brief
stimulation is a muscle twitch. But little if any behavior unfolds on such a short

3
4 The Intelligent Movement Machine

time scale. Neurons in motor cortex are not normally active in 50 ms bursts
but instead, to a first approximation, are active throughout the duration of a
movement. In the present case, the stimulation was applied for half a second,
approximating the duration of a monkey’s reaching or grasping. As a result,
instead of a muscle twitch, a complete movement unfolded.
After a month of mulling and of dinner conversations at the local Italian
restaurant, three of us began the new experiment: Tirin, myself, and Charlotte
Taylor, a graduate student also in the lab. We set out to study the motor cortex
using the technique of stimulating on a behaviorally relevant time scale.
Our procedure was to sit for hours in front of the monkey like a panel of
judges, studying one cortical site in a day, stimulating it hundreds of times
under every condition we could think of, watching every event, discussing
every detail, and arguing over exactly what description to write in the data
book. In addition to our general contributions to the experiment, we each had
specific duties. I was the scribe. Charlotte operated the button that delivered
the stimulation to the cortical site. Tirin fed the monkey a constant supply of
raisins to calm him and entice his arms into a variety of test configurations
(The monkey eventually became obese from the constant snacking.). The study
was intentionally as unstructured and as observational as possible. We did not
know what to expect.
On the first day that we reached the motor cortex it became abundantly
clear that stimulation evoked complex movements combining many joints.
We were able to evoke integrated movements of the shoulder, arm, and hand.
We also noticed that regardless of the starting position of the arm, the move-
ment evoked by stimulation seemed to bring the hand toward the same final
position as if in a goal-directed action.
A few days later we encountered a site in the cortex where stimulation
caused the fingers to close in an apparent grip, the hand to move to the mouth,
and the mouth to open. The monkey appeared to be feeding himself, even though
there was nothing in his hand. The movement was so natural, so utterly like
the monkey’s normal feeding action, that triggering it by button push gave us
the willies. It was uncanny. We ran out of the experiment room and searched the
halls for someone, anyone, to look at the result and tell us that it was real, that
we weren’t nuts.
We wondered if the monkey was inadvertently fooling us. Perhaps the
stimulation caused merely a general tendency to move and the monkey then
supplied a movement that was on his mind, so to speak, because he was con-
stantly feeding himself raisins. This explanation seemed unlikely because we
evoked the hand-to-mouth movement only from one region of cortex, and the
evoked movement had a mechanical reliability. However, we tested the possi-
bility by injecting an anesthetic into the monkey and waiting until he was
asleep. Stimulation of the same site in cortex still drove the fingers into a grip,
the hand upward toward the mouth, and the mouth open. The movement had
nothing to do with the monkey’s behavioral context. It was as mechanical as
clockwork. We appeared to have tapped into its control mechanism.
1. Introduction 5

As the experiment continued, we uncovered more actions that looked like

they were straight out of the monkey’s natural repertoire and that could be
generated by stimulating specific sites in the motor cortex. Different zones
within the motor cortex appeared to emphasize different major categories of
action. Some of these action categories are illustrated in Figure 1-1. They
included ethologically relevant behaviors such as closing the hand in a grip
while bringing the hand to the mouth and opening the mouth; extending the
hand away from the body with the grip opened as if in preparation to grasp
an object; bringing the hand inward to a region just in front of the chest while
shaping the fingers, as if to manipulate an object; squinting the facial muscles
while turning the head sharply to one side and flinging up the arm, as if to

Climbing/leaping

Hand in
Reach to grasp lower space

Manipulate in
central space

Defense

Chewing/
licking

Hand to mouth

Figure 1-1 Action zones in the motor cortex of the monkey. Seven common catego-
ries of movement evoked by electrical stimulation of the cortex on the behaviorally
relevant time scale of 0.5 sec. Images traced from video frames. Each image represents
the final posture obtained at the end of the stimulation-evoked movement. Within
each action zone, movements of a similar behavioral category were evoked. Based on
results from Graziano et al. (2005; Graziano, Taylor, et al., 2002).
6 The Intelligent Movement Machine

protect the face from an impending impact; and moving all four limbs as if
leaping or climbing. The behavioral repertoire of the animal seemed to be
rendered onto the cortical sheet. One might say that the cortical motor system
had an action map.
The evoked movements were also roughly arranged across the cortex accord-
ing to the location in space to which the movement was directed. The height of
the hand was most clearly mapped across the cortical surface. Stimulation of
the lower (ventral) regions of cortex commonly drove the hand into upper space,
and stimulation of upper (dorsal) regions of cortex commonly drove the hand
into lower space (Figure 1-2). Again, an important aspect of the animal’s action
repertoire was mapped across the cortex.
Over the next several years, as I set up my own lab at Princeton, we studied
these cortical action maps with a variety of methods. We measured arm move-
ment at high resolution to better understand the electrically evoked actions.
We chemically activated or inhibited neurons at sites in the cortex and mea-
sured the effect on the monkey’s behavior. We measured the neuronal activity
in motor cortex that occurs during spontaneous movement to determine if the
neurons are naturally tuned to complex actions. We even carried a video camera
to the zoo, and then to an island populated by wild monkeys, to better under-
stand the natural simian movement repertoire.
This line of experiments led us to propose two principles to explain the
basic properties of the motor cortex. One principle concerned the topographic
layout of the motor cortex, and the other concerned the neuronal mechanism
by which motor cortex caused movement.

Topographic Organization
A traditional view of the motor cortex is that it contains a map of the body.
This map was famously depicted by Penfield, whose homunculus diagram is
shown in Figure 1-3. This traditional topographic scheme, however, does not
capture the actual pattern of overlaps, fractures, re-representations, and mul-
tiple areas separated by fuzzy borders. The homonculus does not adequately
describe the topographic organization. A current view of the motor cortex is that
it can be divided into many distinct areas with separate functions (Figure 1-4).
Yet the functions are largely not known, and the properties described thus far
tend to vary across cortex in a graded fashion without hard borders. Rather
than a set of separate areas, the pattern resembles a statistical distribution with
clustering. Labeling those clusters with acronyms, drawing borders around
them, and assigning functions to them may provide a convenient description
but does not explain the principles behind the organization.
Based on our stimulation results, we proposed an underlying topographic
principle for the motor cortex: the reduction of the many-dimensional space
of the animal’s movement repertoire onto the two-dimensional surface of the
cortex. This reduction is similar to the problem in cartography of reducing the
three-dimensional, curved globe onto a two-dimensional map, introducing
unavoidable distortions and fracture lines. In the case of motor cortex, however,
1. Introduction 7

Figure 1-2 Progression of spatial locations to which hand movements are directed.
Within the arm representation of the monkey motor cortex, electrical stimulation in
dorsal cortex tended to drive the hand into lower space; stimulation in ventral cortex
tended to drive the hand into upper space; stimulation in intermediate cortical locations
tended to drive the hand to intermediate heights. Each image is a tracing of the final pos-
ture obtained at the end of a stimulation-evoked movement. Each dotted line shows the
trajectory of the hand during the 0.5-sec stimulation train. Dots show the position of the
hand in 30-ms increments. These trajectories show the convergence of the hand from
disparate starting locations toward a final location. Adapted from Graziano, Taylor, et al.
(2002).

the reduction is from the highly dimensional action space of the animal’s nor-
mal behavioral repertoire to the two-dimensional cortical sheet. The core of
this theory of cortical organization is that local continuity is preserved as much
as possible. Information processors that need to interact are arranged physi-
cally near each other in cortex, presumably gaining a connectional advantage.
One could term this principle of cortical organization the rule of “like attracts
like.” Perfect continuity is not possible, however, because of the unavoidable
difficulties of rendering a highly dimensional space onto a two-dimensional
sheet. The result is a complex compromise among many constraints.
8 The Intelligent Movement Machine

Shoulder
Trunk

Elbow
Hip

Wrist
Kn

d
ee

n
Ha
An
kle

tle
To

M R ing
Lit
es

e
dl
ex

id
Ind mb
u
Th Neck
w
Bro ll
ba
eye
and
lid Face
Eye

V O C A L I Z AT I O N
Lips

Jaw
T
N

Sw ongu
AT I O

allo e
win
SA N

g
LIV
IO
AT

IC
ST
MA

Figure 1-3 The motor homonculus of the human brain from Penfield and Rasmussen
(1950). A coronal slice through the motor cortex is shown. Each point in motor cortex
was electrically stimulated and the evoked muscle twitch was noted. Although each cor-
tical point could activate many muscles, a rough body plan could be discerned.

In our proposal, the map of actions in Figure 1-1 is not by itself correct.
It is present in the data, but the pattern is noisy and approximate. The map of
hand locations shown in Figure 1-2 is also noisy and approximate, and there-
fore not by itself the correct description of motor cortex topography. The map
of the body shown in Figure 1-3 is also present only in a rough sense and does
not capture the complexities of the pattern. The proposal here is that all of
these potential ways to organize movement, and perhaps others, are rendered
onto the cortical sheet simultaneously, resulting in a compromise that does
not neatly follow any single mapping dimension.
To test the validity of this theory of motor cortex organization, we used a
mathematical model that collapsed an approximate description of the monkey’s
1. Introduction 9

3 cingulate motor areas

on medial wall
SMA Central Sulcus
SEF
pre-SMA

or
Primary Mot
PMDr PMDc
F7 F2
FEF

F5 F4
PMVr PMVc

Arcuate Sulcus

Dorsal

Anterior Posterior

Ventral

Figure 1- 4 Some commonly accepted divisions of the cortical motor

system of the monkey. PMDr = dorsal premotor cortex, rostral division,
also sometimes called “Field 7” (F7). PMDc = Dorsal premotor cortex,
caudal division, also sometimes called “Field 2” (F2). PMVr = Ventral pre-
motor cortex, rostral division, also sometimes called “Field 5” (F5). PMVc =
Ventral premotor cortex, caudal division, also sometimes called “Field 4”
(F4). SMA = supplementary motor area. SEF = supplementary eye field, a
part of SMA. Pre-SMA = pre-supplementary motor area. FEF = frontal
eye field.

movement repertoire onto a two-dimensional sheet following the principle of

maximizing local continuity (Aflalo and Graziano, 2006b; Graziano and Aflalo,
2007). The topographic organization generated by the model resembled the
organization of the actual cortical motor system in many respects, including a
rough clustering of movement categories as in Figure 1-1, an approximate
mapping of hand position as in Figure 1-2, the outlines of a body map as in
Figure 1-3, and the outlines of a primary motor area, dorsal and ventral premo-
tor areas, supplementary motor area, frontal eye field, and supplementary eye
field as in Figure 1-4. The theory of a dimensionality reduction was astonishingly
10 The Intelligent Movement Machine

successful in explaining the overarching organization of this large swath of

cortex totaling about 20% of the macaque cortical mantle.

Mechanism of Movement Control

A traditional view of the neuronal machinery of movement control is that
activity at a site in motor cortex propagates down a fixed pathway through the
spinal cord, activating a set of muscles. Based on our stimulation results, how-
ever, the underlying mechanism seems to be less of a simple feed-forward path-
way and more of a network. The effect of the network is to create a specific
class of mapping from the cortex to the muscles, a mapping that can change
continuously on the basis of feedback about the state of the periphery. If the
periphery is relatively still, the mapping from cortex to muscles appears fixed
and resembles the traditional view. But once the state of the periphery is allowed
to vary as in natural movement, the mapping from cortex to muscles becomes
somewhat fluid in a manner that facilitates complex movement control.
For example, when stimulation causes the hand to move to the mouth,
different patterns of muscle activity are generated depending on the starting
position of the limb. If the arm starts to the right of the mouth, stimulation
evokes activity in the shoulder muscles appropriate for pulling the arm toward
the left. If the arm starts to the left of the mouth, stimulation evokes muscle
activity appropriate for pulling the arm toward the right. In effect, the map-
ping from the stimulated site in cortex to the muscles is not fixed. It changes
depending on feedback information about the position of the limb. In this
manner, the network can control limb position.
In general if the network receives feedback information about a specific
variable, such as hand direction, or hand speed, or the posture of the arm,
then the network can learn to control that variable. A network of this type is
not limited to the control of one movement variable. It can in principle control
muscle force directly and also control higher order variables, in combinations
required for the performance of specific actions. A formal neural-network
model that incorporated this principle of “feedback remapping” was able to
control a model arm, successfully generating actions similar to those evoked
in our stimulation experiments.

Theoretical Framework
The computational studies summarized above on topography and mechanism
provide a potential theoretical framework for understanding at least the out-
lines of the motor cortex. In this framework, the purpose of the motor cortex is
to control behaviorally useful actions in the motor repertoire; its complicated
topographic organization is the result of a systematic rendering of the motor
repertoire onto the cortical sheet; and the neuronal pathways between cortical
neurons and muscles are designed to support the multijoint, feedback-dependant
movements common in normal behavior. The goal of the present book is to
elaborate on this theoretical framework.
1. Introduction 11

ORGANIZATION OF THE BOOK

The book is divided into two parts. The first part, ending in Chapter 6, reviews
the previous literature from the discovery of motor cortex to the present,
placing the current proposals into context. Any theory must be able to account
for past results. By the same token, one cannot reject a theory because it fails
to account for a distorted or mythological version of past results. One purpose
of this review, therefore, is to lay to rest some of the common motor cortex
myths, such as the myth of a muscle-by-muscle map. Chapter 6 discusses some
advantages and limitations of the electrical stimulation technique because
much of our work is based on this technique.
The second part, beginning with Chapter 7, describes the experiments and
computational models that form the basis of the present perspective on motor
cortex. Much of this work has been reported piecemeal in published articles.
The present format allows for a more coherent global picture, additional anal-
yses and results, and an extended discussion. Two chapters in particular are at
the heart of the present story. Chapter 10 describes the proposal that the spa-
tial layout of the cortical motor system can be understood as a reduction of
the movement repertoire onto the cortical sheet. Chapter 11 describes the
proposal that the mechanism of movement control by the motor cortex can
be understood as a feedback-remapping mechanism, a divergent mapping
from neurons in cortex to muscles that is continuously remapped based on
information about the changing state of the periphery.
The final chapter of the book discusses possible links between motor control
and social behavior, including the link between defensive movement and social
smiles and between autism and abnormal movement control. The purpose of
this final chapter is to emphasize the point that the motor system is not merely
for activating muscles. It is a machine that allows intelligent interaction with
the environment.

NOTE ON TERMINOLOGY
Figure 1-4 shows a schematic side view of a monkey brain with some com-
monly recognized cortical divisions (e.g., Dum and Strick, 2002; He et al.,
1995; Luppino et al., 1991; Matelli et al., 1985; Matsuzaka et al., 1992; Preuss
et al., 1996; Rizzolatti and Luppino, 2001). The cortical areas directly involved
in motor control are typically divided into a lateral motor strip (unshaded in
the figure) and a medial motor strip (shaded and partly hidden over the crown
of the hemisphere). The lateral motor strip is divided into a posterior strip
termed the “primary motor cortex,” and an anterior strip termed the “lateral
premotor cortex.” The lateral premotor cortex is subdivided into a dorsal pre-
motor area (PMD) and a ventral premotor area (PMV). In the monkey brain,
each of these in turn is subdivided into a rostral area and a caudal area: PMDr,
PMDc, PMVr, PMVc. These areas have also been labeled by Matelli et al.
(1985, 1991) (in the same order) F7, F2, F5, F4. Because different groups have
tended to publish work on different subdivisions, the PMDs are most often
12 The Intelligent Movement Machine

termed “PMDr” and “PMDc,” whereas the PMVs are most often termed “F5”
and “F4.” A region with distinct properties that probably corresponds to the
dorsal-most part of F4 has also been termed the “polysensory zone” (PZ) (Gra-
ziano and Gandhi, 2000). In the human brain, the divisions between rostral
premotor and caudal premotor are less well established and the homology to
the monkey brain is not yet clear.
The medial motor strip (shaded in the figure) was originally labeled the
supplementary motor area, or SMA (Penfield and Welch, 1951). However, this
region has now been subdivided into SMA, pre-SMA that lies directly anterior
to SMA (Matsuzaka et al., 1992), and in the monkey a set of at least three little-
studied areas on the medial part of the hemisphere buried in the cingulate
sulcus, that are termed the “cingulate motor areas” (Dum and Strick, 1991).
Two gaze-control areas are also shown in cross-hatching in Figure 1-4. The
frontal eye field (FEF) lies directly anterior to the arcuate sulcus and in its ante-
rior bank. The supplementary eye field (SEF) lies within the anterior part of
SMA. Both of these gaze areas are defined by the eye and head movements that
can be evoked by electrical stimulation. Eye movements can also be evoked to a
lesser extent from PMDr and PMDc (Bruce et al., 1985; Fujii et al., 2000).
The term premotor cortex is used to refer to at least three different regions.
First, it has sometimes been used to refer to the lateral premotor cortex (PMDr,
PMDc, PMVr, and PMVc). Second, it has been used to refer specifically to the
dorsal part of the lateral premotor cortex (PMDr and PMDc). Third, it has
been used to refer to all cortical motor areas excluding the primary motor
cortex. The looseness with which the term is used can lead to some confusion.
The term motor cortex originally referred to the lateral motor strip, when
that area was believed to be the only motor map of the body. It is now used
variously to indicate the primary motor cortex, the lateral motor strip includ-
ing primary motor and lateral premotor cortex, all cortical motor areas inclu-
sively, or whatever part of the cortical motor system is under discussion at the
moment. Because one theme in this book is that the divisions among motor
areas are not as clear as sometimes suggested, it is useful to have a term that is
intentionally ambiguous.
Chapter 2

Early Experiments on Motor Cortex

INTRODUCTION
This chapter describes how the dominant ideas about motor cortex first emerged.
Many of the forgotten initial observations are still of direct scientific relevance.
Moreover, the history shows how myths and factoids evolved and became resis-
tant to change. Tracing these scientific stories reminds us that the prevailing
beliefs at any time are not to be trusted. Certain beliefs, such as the early view that
the cortex is inexcitable, or the more recent view of a discrete somatotopic map of
the body in the primary motor cortex, are repeated and simplified through repeti-
tion until they become parables of uncertain validity.
This chapter traces motor cortex research from its beginning to the motor
maps of Penfield and Boldrey (1937) and Woolsey et al. (1952). This segment
of the history is mainly about electrical stimulation applied to the surface of
the cortex. Using this technique, researchers drew motor maps of greater and
greater elaboration. After Penfield and Woolsey, more fine-grained techniques
such as microstimulation and single-neuron recording were used to probe the
details and, as might have been expected, reopened all the same questions and
debates. The more modern story of motor cortex, post-1952, is summarized
in Chapters 3 through 5.

SWEDENBORG
There is some variation of opinion about where to begin the history of motor
cortex research. Gross (1997) describes the remarkable case of Emanuel Swe-
denborg, a Swedish philosopher and mystic of the eighteenth century. In 1744
Swedenborg wrote a treatise on the brain. He proposed, among other remark-
ably accurate hypotheses, that movement was controlled by the cerebral cortex;
that the feet were controlled by the uppermost part of the cortex; that the
midsection of the body including the abdomen was controlled by the midre-
gion of the cortex; and that the face was controlled by the lowermost part of
the cortex. At that time the prevailing view of the cerebral cortex was of a
nutritive or protective rind that served no mental function (Gross, 1997), yet
Swedenborg correctly described the functional importance of the cortex and
the upside-down topography of the motor map. Unfortunately his writings
do not describe how he deduced these properties of the cortex. He is known
to have visited contemporary physiology labs and may have observed a set of
suggestive experiments that were never independently published. In any case,

13
14 The Intelligent Movement Machine

Swedenborg appears to have been the first to propose a topographic motor map
in the brain, predating Fritsch and Hitzig by 130 years. His views, however, were
not generally known in his time and had little or no impact (Gross, 1997).

TODD
Reynolds (2004) begins his history of motor cortex with Robert Bentley Todd,
an Irish physiologist who in 1849 published a set of observations and specula-
tions on epilepsy. Todd (1849) attempted to define the brain regions that
caused epilepsy. He argued that only the midbrain and cerebral cortex are
likely to be involved. His evidence for midbrain involvement was that, when
he electrically stimulated the midbrains of rabbits, he evoked general convul-
sions that resembled epilepsy.
On the cerebral cortex, Todd’s observations were less direct. The prevailing
view at that time was that no movement whatsoever could be evoked from the
cortex by any stimulation. It was “inexcitable.” Yet Todd noted that on post-
mortem examination, epileptics were sometimes found to have visible damage
to the cerebral cortex. He described a case study of a two-year-old boy who
developed seizures localized to his left hand. Over several days the seizures
grew worse and spread to the entire left side of his body. On his death shortly
after, the boy was found to have lesions of the dura over the right cerebral
hemisphere. Todd deduced that an irritation or malfunction of the cerebral
cortex on the right side resulted in muscular seizures of the left limbs.
Todd came remarkably close to deducing the motor functions of the cortex.
In his final analysis, however, he shot very wide of the mark. His interpreta-
tion, in hindsight, seems to be an attempt to force his observations into the
prevailing view that the cerebral cortex was inexcitable. Todd suggested a
three-part mechanism: malfunction of the cerebral cortex was responsible
solely for the loss of consciousness and higher mental functions during epilepsy;
malfunction of the midbrain was responsible for the motoric convulsions; and
malfunction of the spinal cord and medulla was uncommon and resulted in
sustained muscle contractions rather than epileptic seizures. If the cortex
played a role in movement control, it was only secondarily by way of the mid-
brain. In his words, “Under ordinary stimulation of the substance of the hemi-
spheres, the fibres are incapable of exiting motion. It is not the office of these
fibres to propagate the nervous force to muscles but to other nervous centres”
(Todd, 1849, p. 999). Although correctly identifying a link between epilepsy
and the cerebral cortex, Todd could not escape the views of his time and there-
fore missed the point of the cortical control of movement.

BROCA
One could almost start the history of motor cortex research with Paul Broca,
the French neurologist who in 1861 described the case study of Tan. Although
Broca met Tan only days before the patient’s death, Broca was able to reconstruct
2. Early Experiments on Motor Cortex 15

some of Tan’s past from interviewing hospital staff and friends. At the age of
thirty one, Tan was admitted to the Hospice of Bicetre with an inability to
speak. In all other respects he was normal and apparently intelligent, but other
than the word “tan-tan” (from which he acquired his nick name) and an unre-
corded gross swear word that he used when frustrated, he had lost the ability
to produce vocal language. Over the next twenty-one years he slowly devel-
oped a weakness and then a paralysis of the right arm, followed by a similar
paralysis of the right leg. He became bedridden. Because his sheets were
changed only once a week, he developed an infection of the right leg that was
not noticed until after it had become life threatening. It was at this point that
Broca examined him. A few days later the patient died, and Broca performed
an autopsy of the brain. The left frontal lobe was badly degenerated with a
focus of degeneration in the third frontal convolution, now commonly known
as Broca’s area.
Broca concluded that the lesion must have begun small, affecting the specific
cortical center for speech, and then gradually spread to surrounding tissue,
including the precentral gyrus (now known to be the site of motor cortex).
This spread of the degeneration, according to Broca, must have caused the
gradual paralysis of the body. Broca, like Todd, came within a hair’s breadth of
deducing the motor map. His detailed observations placed him in exactly the
right part of cortex. His careful estimates of the center of the degeneration
and its rate of spread could have led him to a cortical map arranged sequen-
tially from one body part to the next, from face to arm to leg, in the order of
the progression of Tan’s symptoms. He was willing to infer that at least one
function, speech, was localized to a region of the cerebral cortex. Yet in his
words, “Everybody knows that the cerebral convolutions are not motor organs.
The corpus striatum of the left hemisphere is of all the attacked organs the
only one where one could look for the cause of the paralysis” (Broca,
1861/1960, p. 70). Because of his acceptance of the beliefs of the time, he was
totally unable to see the importance of his observations for motor control.
Surely the lesson here is to be most wary of the thing that “everybody knows.”

JACKSON
Most historical reviews of motor cortex research begin with the English neu-
rologist John Houghlings Jackson (e.g., Ferrier, 1873; Foerster, 1936; Hitzig,
1900; Penfield and Boldrey, 1937). Jackson is generally credited with having
deduced the existence of a somatotopic motor map in the cortex on the basis
of the spread of epileptic seizures across the body. A close reading of his work,
however, shows that this common belief about Jackson is completely wrong. It
is an interesting case study in the way that historical myths become estab-
lished in science. Jackson is something of the Nostradamus of neuroscience;
his writing is ambiguous and rich enough that one can read almost anything
into it. For the sake of getting the story right, it seems worth detailing the
frankly brilliant ideas that Jackson actually did propose.
16 The Intelligent Movement Machine

During the 1860s Jackson studied a large number of epileptic cases that he
summarized in a publication in 1870. Some of his patients suffered from
global seizures that simultaneously affected the entire body. Others suffered
from partial seizures that began in one location on one side of the body. Jack-
son focused his theoretical work on the cases of partial seizures. In his writing
he was quite clear that the seizures were caused by malfunction of the cerebral
hemispheres, but he was inconsistent on whether they were caused by mal-
function of the cortex or of the striatum, a large nucleus underlying the cor-
tex. Jackson implied that the partial seizures, being simpler, were probably
caused by the striatum, whereas the global seizures, being more complex, may
have been caused by the cortex. In this respect Jackson fell into the same trap
as Todd (1849) and Broca (1861/1960). In every case of a partial seizure in
which he was able to examine the brain afterward, Jackson described damage
to the cerebral cortex, not to the striatum; yet he seemed unable to let go of
the idea that the cortex was too complex a structure for the control of a body
part.
Jackson (1870) noticed that the partial seizures almost always began in the
hand, in the face around the mouth, or in the foot, the three parts of the body
that are most commonly used, or in his description, that have the most “varied
uses.” Furthermore, “fits which begin in the hand begin usually in the index
finger and thumb; fits which begin in the foot begin usually in the great toe”
(p.10). From these observations he deduced that the amount of neurons in
the brain devoted to a body part, and therefore the chances that the body part
may be affected by seizures, must be proportional to the amount of use of the
body part. Jackson therefore brilliantly formulated the general principle of
brain organization that behaviorally important functions have physically larger
representations.
Jackson also noticed that seizures beginning on the right side of the face
and tongue were often followed by a lingering loss of speech. He inferred that
these right-sided facial seizures were caused by nervous instability and explo-
sive discharge in Broca’s recently described speech area in the left hemisphere
(Broca, 1861/1960). After the discharge, the brain area must suffer from fatigue
resulting in a loss of speech. Jackson further deduced that because seizures
beginning in the hand and the foot did not affect speech, they must be caused
by instability of other, separate brain regions. He therefore correctly deduced
that the brain contained different centers for the control of different body
parts.
Jackson (1870) noticed that, “When a fit begins in the hand it goes up the
arm and down the leg . . . Now patients who have fits beginning in the foot
tell me that the spasm goes up the leg and down the arm” (p. 23). This pro-
gression of convulsions from one body part to the next is now known as a
Jacksonian march. This observation was not news to the patients who com-
monly tried to block the march by tying ligatures around their limbs. One
woman with a seizure that always began in the hand would tie a ligature
around her wrist; and when that did not work, she tied it higher up the arm.
2. Early Experiments on Motor Cortex 17

Another patient with seizures that began in the foot similarly tied a ligature
around his ankle.
To Jackson, this spread of epilepsy from one body part to the next pre-
sented a theoretical problem. He believed that a partial seizure was caused by
abnormal instability in a focused spot in the brain. In his words, “The fact
that the symptoms are local implies, I hold, that there is of necessity a local
lesion” (Jackson, 1870, p. 24). Yet a seizure that starts in the hand may spread
to other body parts and in some cases may spread to the entire body bilater-
ally. Therefore, to Jackson, all the parts of the body affected by the seizure
must be represented within the local, diseased brain region. He asks, “Why, if
face, arm and leg are represented together in the square inch, is the fit a sequence
only? Why are not all these parts convulsed contemporaneously?” (p. 27). His
answer is that the diseased “square inch” of the brain must normally control
useful sequences of actions. During a diseased discharge, the sequence of
spasms or the spread across the body is a crude caricature of the stored
sequence of actions.
Even by 1875, five years after Fritsch and Hitzig published their physiologi-
cal map of motor cortex, Jackson (1875) wrote unambiguously: “When we
grasp . . . the more strongly the hand is used, the farther up the arm does the
movement spread” and therefore, “if a fit begins in the thumb and index
finger, there will probably be developed . . . that series of movements which in
health serves subordinately when the thumb and index finger are used”
(p. 69). Jackson, therefore, failed totally to appreciate the true reason for the
progression of partial seizures across the body, namely the spread of an epilep-
tic storm across a map of the body in the cortex. He accepted the lay belief
that tying a ligature around the limb can stop the spread of epilepsy up the
limb, a view that is totally untenable if the spread is actually across the cortex.
Jackson had no concept of the spread of epilepsy across the cortex, nor had he
any concept of a somatotopic map in the cortex.
The mechanism of motor control that Jackson deduced was a collection of
centers, possibly in the striatum, each one of which controlled the entire body
with an emphasis on coordinating the action of one particular body part.
Centers emphasizing the hand, face, and foot were larger than centers empha-
sizing other body parts because the hand, face, and foot required a more
complex and varied movement repertoire. In these deductions he came
remarkably close to the truth, but not as close as is sometimes suggested.
Jackson saw many of his speculations confirmed in 1870, when Fritsch and
Hitzig stimulated the dog brain and demonstrated a set of distinct centers that
corresponded to different parts of the dog’s body. These centers were located
in the cerebral cortex rather than in the striatum; but aside from his localiza-
tion error, Jackson’s essential concepts appeared to have been vindicated.
As physiological evidence for his movement centers accumulated, Jackson
developed an overarching description of the brain basis of behavior. In 1890
he proposed that the control of movement could be divided into three levels
that corresponded to three stages in animal evolution (based on a flawed
18 The Intelligent Movement Machine

understanding of Darwin’s theory of evolution typical of that time, in which

evolution progressed from lower to higher levels). The lowest level of motor
control was subcortical, controlling the simplest elements of movement. The
middle level lay in the motor cortex of Fritsch and Hitzig, controlling move-
ment in an integrated and complex fashion. The highest level lay in the pre-
frontal cortex, controlling movement in the most abstract sense of will or
thought. Epileptic seizures of the highest level caused a loss of consciousness;
seizures at the middle level caused muscular convulsions; seizures at the
lowest level normally did not occur and artificial stimulation here caused
sustained contractions. It is only fair to point out that this influential idea
indelibly associated with Jackson’s name is not entirely original to Jackson.
It is an elaboration and modification of Todd’s 1849 proposal. Although Todd
did not express the idea as clearly and attributed his functions to a slightly dif-
ferent underlying anatomy, he outlined a highest level of consciousness that
lay in the cortex (seizures here induced loss of consciousness), a middle level
of movement control that lay in the midbrain (seizures here induced muscle
convulsions), and a lowest level of movement control that lay in the spinal
cord and medulla (normally seizures did not occur here and direct stimula-
tion caused sustained muscle contraction).
Jackson was particularly clear that none of the three levels of movement
control, not even the lowest level, was a simple activator of muscles. He wrote:
“motor centers of every level represent movements of muscles, not muscles in
their individual character” (Jackson, 1890, p. 687). By 1890, therefore, the
“muscles vs. movements” debate of motor control was fully framed.

DISCOVERY OF MOTOR CORTEX: FRITSCH, HITZIG, AND FERRIER

In 1870, the two German scientists Fritsch and Hitzig published a set of exper-
iments on the cortex of the dog brain. Borrowing Frau Fritsch’s dressing table
as an operating surface, they exposed the brains of dogs (unanesthetized and
anesthetized; the results were similar) and stimulated the cortex using brief elec-
tric discharges from a battery. In the anterior region of cortex, this stimulation
evoked muscle twitches. Fritsch and Hitzig described distinct cortical centers
from which twitches of different body parts could be obtained. Figure 2-1
shows a redrawing of their summary map of cortical centers. They described
five centers, indicated in the figure by the five cortical sites at which optimal
results could be obtained. For example, stimulation at one site evoked move-
ments of the foreleg; stimulation near the site evoked these movements less
well; stimulation of sites between the labeled points evoked either no move-
ment or, if the electrical current was turned up, combined movements of the
adjacent centers. In two animals, Fritsch and Hitzig removed part of the fore-
leg center on the right side and observed that the dogs were thereafter unable
to move the left foreleg in a coordinated fashion while walking or running.
Fritsch and Hitzig may have been steeped in the beliefs of the time, but
unlike previous investigators of movement control, they took their data at face
2. Early Experiments on Motor Cortex 19

Neck

Foreleg

Hindleg

Face

Figure 2-1 The map of stimulation-evoked

movements in a dog brain adapted from Fritsch
and Hitzig (1870/1960). Each point indicates
the approximate location of a movement center.
Stimulation at or near each point evoked move-
ments of the indicated body part.

value. They drew two broad conclusions. First, contrary to the entrenched
belief, the cerebral cortex was in fact “excitable” in that stimulation of it caused
overt behavior. Second, contrary to Flourens’ notion (1824/1960) of a homog-
enous cerebral cortex (at least in a chicken, Flourens’ animal of choice), in
which all functions were equally distributed, Fritsch and Hitzig (1870/1960)
found that the front or anterior half of the dog cortex “stands in immediate
connection to muscular movements” while the back or posterior half “has evi-
dently nothing to do with it” (p. 92).
Note that Fritsch and Hitzig did not describe a somatotopic map of the
body in the modern sense. Beginning students of neuroscience today are taught
that there is an elaborate roster of body parts separated and placed in order
along the cortex. The map of Fritsch and Hitzig, however, is not so precise.
First, the “map” is not continuous but instead is divided into five islands each
one surrounded by relatively inexcitable cortex. Second, no clear topography
is reported within each island; instead each island represents an undifferenti-
ated collection of many muscles. Third, the rough overall arrangement is not
topologically correct. The neck is represented at the anterior end of the “map”
and the face is represented at the posterior end behind the hind leg.
Fritsch and Hitzig came down clearly on the side of the cortical control of
movements rather than muscles. In their view, a cortical center was “a middle-
man . . . in which a similar but better coordination of muscle movement takes
20 The Intelligent Movement Machine

place than in the gray substance of the spinal cord or brain stem” (Fritsch and
Hitzig, 1870/1960, p. 92). Furthermore, they understood that movement con-
trol cannot be separated from sensory processing. After lesions of the foreleg
center, the dogs not only lost the ability to control coordinated movements of
the leg but also “had obviously only an imperfect consciousness of the shape
of their member, they had lost the faculty to make a complete conception of
it” (Fritsch and Hitzig, 1870/1960, p. 96). This view of highly complex cortical
centers that controlled movement by combining sensory and motor process-
ing closely resembled Jackson’s concept of motor centers, although it appears
that Fritch and Hitzig were unaware of Jackson’s work at the time.
Shortly after Fritsch and Hitzig’s initial discovery of motor cortex in the
dog brain, Ferrier (1873) replicated the results and extended them from dogs
to cats, rabbits, and guinea pigs, though he was unable to evoke movements
from the brains of birds, complaining that they were too soft and oozed when
stimulated. He went on to study monkeys (Ferrier, 1874), establishing the
motor map on the precentral gyrus of the primate brain with the leg repre-
sented in a dorsal location and the mouth represented in a ventral location.
Figure 2-2 shows a redrawing of Ferrier’s motor map with seven cortical
regions arranged along the precentral gyrus, numbered as he originally num-
bered them. In addition to mapping the motor cortex along the precentral
gyrus, Ferrier also evoked eye and head movements from cortical regions that

2
5 3

6
7
8
9

Figure 2-2 The map of stimulation-evoked movements in a

monkey brain adapted from Ferrier (1874). The cortical regions
arranged along the precentral gyrus are numbered as he origi-
nally numbered them. The effects of stimulation were: Region 2,
leg and foot movement. Region 3, trunk, leg, and tail movement.
Region 5, combined arm and leg movement. Region 6, arm and
hand movement, including some hand-to-mouth movements.
Regions 7–9, mouth movement.
2. Early Experiments on Motor Cortex 21

became known as the frontal eye field (FEF) and the posterior eye field (now
usually known as the lateral intraparietal area or LIP, though Ferrier was not
as specific about its location).
Ferrier’s experiments differed from Fritsch and Hitzig’s in several impor-
tant respects. First, one must credit Ferrier with the first true description of a
motor map in the modern sense of map. He described the somatotopic pro-
gression in detail.
Second, rather than use a direct current pulse from a battery as Fritsch and
Hitzig had done, Ferrier used an alternating current that could be extended
over several seconds. By extending the stimulation, he found that the previ-
ously described muscle twitches unfolded into longer, apparently coordinated
actions that he called purposive movements. “The movements . . . resulting
from excitation of the individual centers are purposive or expressional in
character, and as such we should, from psychological analysis, attribute to ide-
ation and volition if we saw them performed by others” (Ferrier, 1873, p. 73).
For example, in one experiment, stimulation within the center for control of
the leg (circle 2 in Figure 2-2) caused an action “just such as when a monkey
scratches its abdomen with its hind leg” (Ferrier, 1874, p. 413). In another
experiment, stimulation within the center for the hand and arm (circle 6 in
Figure 2-2) “brings the hand up to the mouth, and at the same time the angle
of the mouth is retracted and elevated” in a manner resembling a feeding
movement (Ferrier, 1874, p. 418).
Ferrier argued with Fritsch and Hitzig over the correct method of stimula-
tion. In Ferrier’s view (1873, 1874), the shorter stimulation missed essential
movements, whereas in Hitzig’s view (1900) the longer stimulation evoked
seizures and therefore produced artifactual results. Despite the squabble, their
views on motor cortex were almost identical. They both believed that motor
cortex contained a set of cortical centers that controlled movement at a high
level, coordinating groups of muscles in a meaningful fashion rather than
controlling individual muscles.
A lecture that Hitzig published in English in 1900, thirty years after the initial
discovery of motor cortex, is a vivid editorial on his contemporaries and reveals
something of the issues surrounding the research at that time. He is mainly
complimentary of Jackson of whom he predicts, “His thoughts will ever again
rise from the seemingly lifeless dust, and will spur posterity on to renewed
intellectual labour in the field that he cultivated” (p. 546). He states that Goltz
had “given occasion to unpleasant conflicts and thereby also to the spreading
of great confusion over the questions with which we are busied” (p. 558).
Goltz had suggested that Fritsch and Hitzig’s (1870/1960) findings were caused
by the spread of electric current from the surface of the cortex to the underly-
ing striatum. Hitzig argues forcefully against Ferrier’s view that the motor cor-
tex is dedicated to movement control and does not participate in sensation.
He argues equally forcefully against Schiff ’s view that the motor cortex is
purely a sensory area with no true motor function. He systematically antago-
nizes every one of his contemporaries and then ends the address with the
22 The Intelligent Movement Machine

uplifting platitude that scientists form “an army that knows no separation into
different camps . . . in the battle against ignorance” (Hitzig, 1900, p. 581).

ELABORATION OF THE MOTOR MAP: BEEVOR AND HORSLEY

Beevor and Horsley (1887, 1890) conducted a series of experiments that
greatly extended the findings on motor cortex but also, in some ways, retreated
from the initial insights of Fritsch and Hitzig (1870/1960) and of Ferrier
(1874).
In 1887 Beevor and Horsley published a study focusing on the arm and
hand representation in the monkey motor cortex. The study built on Ferrier’s
(1874) work in that it examined the complex or “purposive” movements
evoked from the cortex. The authors outlined a cortical map that was not
merely a plan of the body’s muscles but was an arrangement of useful move-
ments. I find this study particularly interesting because it predicts many of
our own observations one-hundred-and-ten years later (Graziano et al., 2005;
Graziano, Taylor, et al., 2002). Beevor and Horsley’s study may be the last to
examine the topographic mapping of complex actions on the motor cortex
until ours.
In their experiment, Beevor and Horsley found that stimulation of the
lower or ventral part of the arm representation tended to evoke movements
consistent with bringing an object toward the body, especially toward the
mouth for feeding. These movements included flexion of the elbow, supina-
tion of the forearm (orienting the palm toward the body), and closing of the
grip. These authors therefore roughly located the hand-to-mouth zone that
was redescribed much later (Graziano, Taylor, et al., 2002). Because the hand-
to-mouth zone can be several millimeters wide, it is reasonable that Beevor
and Horsley were able to localize it using their relatively coarse surface stimu-
lation through two electrical poles spaced 2 millimeters apart.
Beevor and Horsley also found that stimulation of the upper or dorsal part
of the arm representation tended to evoke “advancing” movements of the arm
including a pronation of the forearm and opening of the hand, which they
variously interpreted as “defensive” or, in a later paper that extended the same
work (Beevor, 1888), as “reaching forward . . . to seize” an object. They described
a specific, circumscribed region of the precentral gyrus that, when stimulated,
consistently evoked this open-hand arm-extended movement. This region is a
close match to our reach-to-grasp zone from which we evoked precisely the
same movement (Graziano et al., 2005). It is also a close match to the modern
caudal dorsal premotor cortex (PMDc) thought to play a role in reaching.
In addition to studying the complex actions evoked at each cortical site,
Beevor and Horsley also studied what they termed the “primary” movement,
the initial or at least the most visible twitch to occur with a short pulse of
stimulation. They noted that the cortical map of primary movements followed
the approximate plan of the body, whereas the cortical map of purposive
actions, or of the primary, secondary, and tertiary movements evoked with
2. Early Experiments on Motor Cortex 23

longer stimulation, was much more overlapping. In their discussion they

imply that the two types of map, a somatotopic map of the body and a map of
complex movements, do not negate each other, and that both are useful ways
to view the cortical organization.
In their later experiments, however, Beevor and Horsley confined their
investigations to the primary movements, perhaps because they were easier to
study or presented fewer interpretational complexities. In 1890 they published
a meticulous study in which they dissected away the cortex overlying the
motor area and stimulated the severed ends of fibers. In this manner they
showed that the movements evoked from the motor cortex depended on the
fibers of the pyramidal tract that originate in cortex and course down through
the internal capsule, the midbrain, and the medulla to the spinal cord. On
stimulating these fibers they obtained a detailed map of the body that pre-
cisely matched the map in the overlying cortex. Their map of motor cortex is
redrawn in Figure 2-3.
Beevor and Horsley’s 1890 experiment represented a move toward a “mus-
cles” view of motor cortex function and away from a “movements” view. This
study suggested that the movements evoked by stimulating the cortex were
caused by the cables that ran down to the spinal cord; the motor cortex itself
was merely the start point of these cables; and the intrinsic processing in the
cortex was of minimal importance because the same map was obtained
whether one briefly stimulated the cortex or the cables. Beevor and Horsley’s
map (Figure 2-3) was a detailed stacking of body parts. In some parts of the
map a single joint such as the shoulder, elbow, or wrist was assigned a discrete
area. This view of motor cortex was radically different from the view presented
in their earlier (Beevor and Horsley, 1887) paper or proposed by Ferrier (1874)

Knee Hallux
Hip Ankle
Sm
Head all
to Shoulder To
es
Opposite Elbow
Side Wrist
Synch Fingers
Head r:
& INDEX
to Op Eyes Opp
p-Side Ang Eye LidsTHUMB
Ey O f l e
e ElevMouth
t s atio
Op o n
S p: outh
& Mide Angle of M n)
etractio
icat:
Pharnx

Lin d i (R
e
x

Mast

Opening
Laryn

of Mouth

Figure 2-3 The map of stimulation-evoked movements in the monkey brain from
Beevor and Horsley (1890).
24 The Intelligent Movement Machine

or Fritsch and Hitzig (1870/1960). In the earlier view, motor cortex was a col-
lection of motor centers. Each center was a sophisticated processing device
that produced meaningful actions by intelligently combining the movements
of many muscles and joints and perhaps also by integrating sensory input
with motor output. Beevor and Horsley’s 1890 paper instead implied a view in
which there were no motor centers, and in which the function of motor cor-
tex was defined by its descending connections to the spinal cord—an antinet-
work, feed-forward view that, unfortunately, came to dominate the twentieth-
century research on motor cortex and to some extent is still dominant today.

ELABORATION OF THE MOTOR MAP: SHERRINGTON

In 1901 Sherrington and his students began a series of unusually varied and
revealing experiments on the motor cortex of anesthetized apes (e.g., Grunbaum
and Sherrington, 1901, 1903; summarized in Sherrington, 1939). They electri-
cally stimulated the motor cortex of twenty-two chimpanzees, three gorillas,
and three orangutans.
To obtain as discrete a map as possible, Sherrington used surface stimulation
that was as weak and brief as possible while still evoking movements. In this
way Sherrington continued a trend away from studying the representations of
complex actions. His map of a chimpanzee’s motor cortex is shown in Figure
2-4. He suggested that this body plan on the cortex closely follows the plan of
the spinal roots that innervate the muscles. He also suggested that the map is
more discrete in “higher types” or apes, less discrete in monkeys, and even less
so in dogs, because apes have the greater need to combine separate move-
ments into novel combinations. (Sherrington, like most of his contemporaries
and like Jackson before him, had not properly grasped Darwin’s concept of
evolution and therefore wrote about lower and higher animals as well as lower
and higher levels of evolution.)
Yet Sherrington was too accurate an observer to reduce the cortex entirely
into a map of muscles. He emphasized the complexity of cortex and its poten-
tial for information processing through the interaction between cortical
points. In particular, he noticed that stimulation of one site in cortex could
alter the movement evoked from another site. He described several types of
interaction in which the movement evoked from a site could be modified,
changed from flexion to extension, or even changed to a different body part
entirely, depending on the prior stimulation of other sites. The map was there-
fore not anatomically fixed but instead highly labile and subject to experience.
He arrived at the conclusion that the motor cortex is not merely a device for
the descending control of the muscles; a major part of its function was the lat-
eral linking of different sites into complex patterns. In his words, “a property
possessed by the motor cortex is the combining of a large, though exhaustible,
number of movements . . . into sequences of very great variety” (Sherrington,
1939, p. 424). This view is similar to a view more recently suggested by Huntley
and Jones (1991) and Schneider et al. (2002).
2. Early Experiments on Motor Cortex 25

Anus & vagina

Sulcus
Toes centralis
Ankle Abdomen
Knee Chest
Hip

Shoulder
Elbow Body
Wrist

Eye
Fingers

s
& thumb

Neck
Eyes
Ton outh
M
gu
e

Ear
Eyelid Closure
Nose of jaw Sulcus centralis
Opening Vocal Mastication
of jaw cords

Figure 2- 4 The map of stimulation-evoked movements in a chimpanzee brain from

Grunbaum and Sherrington (1901). The stippled strip indicates the motor cortex. The
stippled circular area represents the frontal eye field.

Sherrington made several other important observations on the motor cortex

of apes. He noted that the posterior border of the excitable zone, the zone
from which movements could be evoked, was sharply defined; but the ante-
rior border on the precentral gyrus was more of a gradient than an edge. The
anterior border was not only gradual but also unstable. If he stimulated at the
posterior edge of motor cortex and marched the electrode forward in a series
of stimulations until no more movement was evoked, he obtained a relatively
anterior border and a relatively wide expanse of motor cortex. If he reversed
the series, stimulating first in an anterior region outside of motor cortex and
marching the electrode backward until a movement was evoked, he obtained a
different border, a more posterior one that resulted in a narrow expanse of
motor cortex. This observation is important given the later attempts to divide
the precentral gyrus into a posterior, primary motor strip and an anterior,
premotor strip. The physiological properties, at least as seen at this time, did
not suggest a clear border between two different cortical motor areas. The dif-
ferences took the form of a labile gradation, not a categorical distinction.
Finally, Sherrington (1939) performed a set of unusually revealing lesion
experiments. After removing a region of motor cortex, he observed the specific
26 The Intelligent Movement Machine

loss of motor function and how the loss changed over the subsequent days.
For example, after he removed the hand representation on the left side of a
chimp brain, the chimp’s right hand was weak and no longer coordinated. The
chimp seemed to know what it wanted to do but couldn’t perform the move-
ments. In Sherrington’s interpretation, “Surprise at the failure of the limb to
execute what it intended seemed to be the animal’s mental attitude” (p. 435).
Sherrington therefore suggested that the lesion must have caused a “defect in
the motor execution rather than in the mental execution of the act” (p. 435).
The intention to make a movement, in his speculation, was not contained
within motor cortex; it must be localized somewhere else in the brain, perhaps
in the parietal lobe. This insightful speculation foreshadows the much more
recent work of Andersen and colleagues (e.g., Snyder et al., 1997) on move-
ment intention represented in the parietal lobe.
Within a month after the lesion, the chimp had recovered. It was able to
move its right hand in a coordinated fashion. Some other brain area must
have taken over the control of the hand. In a series of secondary lesions, Sher-
rington was not able to find the cortical region responsible for the regained
function. Lesioning the cortex directly around the original lesion, the more dis-
tant cortex representing other body parts, or the hand representation in the
opposite hemisphere, did not take away the regained function in the right hand.
This result suggested that the motor cortex as understood at the time must not
be the only area of cortex capable of controlling voluntary movement. In effect,
Sherrington had provided evidence that the cortex probably contained more
than one motor area, and that the others were yet to be discovered.

CAMPBELL AND THE PROPOSAL OF TWO DISTINCT MOTOR MAPS

Originally the term motor cortex referred to the motor map of the body along
the precentral gyrus. There was no other motor map and no need for a more
precise term. In 1905, Campbell published a division of cortex on the basis of
the appearance of cells and fiber tracts. He suggested that the motor cortex
was divided into two cytoarchitectonic regions: a posterior region that was
characterized by a dense population of giant pyramidal-shaped cells (termed
“Betz cells”) in the deeper layers, and an anterior region that was similar in
most respects but lacked the Betz cells. Campbell acknowledged that the two
areas do not have an abrupt border. They grade into each other. Nonetheless
he argued for a distinction between them. (One might say that cytoarchitec-
tonics is the taking advantage of the human perceptual tendency to see bor-
ders where a gradient actually exists.) The posterior strip he termed the “pre-
central” cortex and the anterior strip he termed the “intermediate precentral”
cortex (Figure 2-5). Brodmann (1909) saw a similar division of the motor
cortex into two areas that he termed “area 4” (Campbell’s precentral cortex)
and “area 6” (roughly matching Campbell’s intermediate precentral cortex).
Today Brodmann’s terminology is more common.
2. Early Experiments on Motor Cortex 27

Intermediate precentral
cortex Precentral
cortex

Figure 2-5 The division of the human motor cortex into a posterior strip, the
“precentral” cortex, and an anterior strip, the “intermediate precentral” cortex,
from Campbell (1905).

Campbell invoked the three motor levels of Jackson (1890). The lowest
level was in the spinal cord and medulla. The middle level, according to
Campbell’s proposal, was the precentral strip characterized by Betz cells.
Campbell viewed this region of cortex as the true or “primary” motor cortex in
direct control of muscles. In this way Campbell officialized the term primary
motor cortex, assigning Beevor and Horsely’s (1887) “primary” movements to a
specific sector of cortex. Finally, Campbell placed the highest level of motor
control, involved in the coordination of complex actions, at least partly in his
proposed intermediate precentral cortex.
Campbell amassed a set of reasons to support this division of motor cortex
into two motor areas, a “primary” one and an “intermediate precentral” one.
Campbell first argued that the giant Betz cells were the critical output cables
from the cortex controlling movement. This speculation turned out to be
incorrect. The Betz cells compose only about 3% of the descending projection
from the cortex to the spinal cord in primates (e.g., Lassek, 1941). Yet to Camp-
bell the Betz cells were the primary conduit by which the cortex controlled the
body. He examined the brains of patients who had suffered progressive loss of
muscle control, and he detected no abnormality except a loss of giant Betz
cells in motor cortex. He studied the brains of amputees and discovered a loss
of Betz cells in the region of motor cortex corresponding to the missing limb,
as if, lacking muscles to control, the Betz cells had died. Indeed Campbell
rather brilliantly demonstrated the map of the body in the human motor cor-
tex by examining the patterns of degeneration caused by different amputations.
28 The Intelligent Movement Machine

Because in Campbell’s (incorrect) view the Betz cells formed the main or only
motor output from cortex, and because according to Campbell the Betz cells
were found essentially in the posterior motor strip and not the anterior strip
(a view that was also not quite correct; the distinction is more a gradient than
a border; e.g., Bucy, 1935), therefore only the posterior strip was the true or
primary motor area whereas the anterior strip must serve a different function.
Campbell’s arguments for the complex functions of the anterior strip, the
intermediate precentral cortex, were equally speculative. Broca’s area was clearly
responsible for the high-level organizing of movements into speech because
lesions to it eliminated speech without eliminating the raw ability to move the
mouth muscles (Broca, 1861/1960). The intermediate precentral cortex con-
tained Broca’s area, therefore the intermediate precentral cortex must generally
control the complex coordination of movement. Note, however, that Camp-
bell’s intermediate precentral cortex (Figure 2-5) looks almost as though it has
been gerrymandered to include Broca’s area. Its ventral part extends like a foot
with the toes in Broca’s third frontal convolution.
Agraphia, a hypothetical syndrome in which a patient is selectively unable
to write, Campbell speculatively localized to a region of the intermediate pre-
central cortex just anterior to the primary motor hand area, a region now
thought to be mainly an eye movement area unrelated to hand movement.
Campbell also speculated that a high-order leg area, controlling complex
actions of the leg, must be located just anterior to the primary motor leg area,
in a region now typically considered to be the supplementary motor cortex.
The specifics of Campbell’s hypothesis are therefore almost all wrong.
Campbell speculated beyond his data and arrived at a sequence of errors.
Perhaps his fundamental error was an overreliance on the dubious doctrine that
the function of a brain area can be deduced from its appearance under a micro-
scope. Yet his two main suggestions were influential. He suggested first that the
motor cortex could be divided into an anterior and a posterior area; and second,
that the areas were hierarchically linked, the anterior one controlling the poste-
rior one, which in turn controlled the spinal cord. These suggestions continue
to resonate today.

VOGT AND VOGT AND THE ELABORATION OF

THE MOTOR HIERARCHY
After Campbell, a range of experiments supported the hypothesis that the pre-
central gyrus was not uniform. For example, the German team Cecile Vogt and
Oskar Vogt (1919, 1926) divided the monkey motor cortex into a large number
of cortical fields on the basis of appearance under a microscope. Their divisions
included a primary motor field and several ventral and dorsal premotor fields.
Most of their physiological work focused on the dorsal motor areas, shown in
Figure 2-6. Their first cortical field, area 4, was the most posterior region and
was dense with Betz cells. They referred to this area as the primary motor area
because stimulation at low currents evoked simple movement twitches of
2. Early Experiments on Motor Cortex 29

6a 4
6a

Figure 2-6 Division of the monkey motor cortex into three

fields, including two dorsal premotor fields, adapted from
Vogt and Vogt (1926). The scheme of Vogt and Vogt incorpo-
rated several other divisions including a set of ventral premo-
tor areas, but their physiological work focused on the dorsal
areas shown here.

separate body parts. The secondary field, 6aα, was just anterior to the primary
motor field. Stimulation here at low currents evoked simple twitches similar
to those evoked from primary motor cortex. However, stimulation at higher
currents evoked more complex movements that combined more than one
body part. The tertiary field, 6aβ, was just anterior to the secondary field.
Stimulation here at low currents evoked no movement at all whereas stimula-
tion at higher currents evoked complex movements similar to those evoked
from 6aα. To the Vogts, these borders between areas were absolute and so precise
that they were “hairline” divisions (Vogt and Vogt, 1926). In this respect of
hard borders, the Vogts represented an extreme view. Other researchers saw at
least some gradation rather than hard borders between cortical fields (e.g.,
Broadman, 1909; Bucy, 1935; Campbell, 1905).
To further understand the possible sequence of processing among these
motor fields, Vogt and Vogt (1919) performed a set of experiments combining
stimulation with fiber cutting. These experiments were not described in great
detail in published form and were repeated and extended by Bucy (1933), and
therefore Bucy’s experiments are described here. Bucy did not distinguish
between 6aα and 6aβ. His experiments involved a comparison between area 6
and the primary motor cortex, area 4. Cutting the cortex between areas 6 and 4
caused little or no effect on the movements evoked from surface stimulation
of area 4. The cut, however, abolished the simple movements evoked from
area 6 while leaving intact the more complex movements evoked at higher
currents. Complete removal of area 4 also abolished the simple movements
evoked from area 6 while leaving intact the more complex movements evoked
at higher currents. In exact contrast, cutting the white matter beneath area 6,
30 The Intelligent Movement Machine

thus disconnecting its deep projections, eliminated the complex movements

while leaving the simpler movements evoked from area 6. These results sug-
gested that stimulation of area 6 evoked movements by means of two different
mechanisms: the simpler movements depended on lateral connections passing
into area 4, fibers that could be severed with a cut across the cortex; and the
more complex movements depended on deeper connections perhaps directly
to the spinal cord or to subcortical nuclei. Whether to describe 6 and 4 as
linked in a hierarchical series, or whether to describe them as hierarchically
equal and operating in parallel though processing different aspects of move-
ment, was therefore ambiguous. One could find evidence of both types of
relationships.

DEFINING PREMOTOR CORTEX: FULTON

It is not clear who coined the term premotor cortex to refer to the hypothesized
higher-order motor cortex. Hines (1929) used the term in passing, but he may
not have been the first. Given that prefrontal was already the accepted term for
the anterior part of the frontal lobe, perhaps it was natural to use premotor to
refer to the anterior part of motor cortex. In any case, the term is usually associ-
ated with Fulton who gave the name its modern connotation and popularized it
(Fulton, 1934, 1935). Although Fulton performed stimulation and anatomical
experiments, it was largely on the basis of lesion work in monkeys that he elabo-
rated his concept of a premotor cortex. His starting point was an observation
made by Richter and Hines in 1932.
Richter and Hines (1932) found that lesions to area 6 in monkeys caused
uncontrollable grasping in the opposite hand. A monkey could even be hung
from a bar by his grip. The dysfunction was temporary; after a few days the
monkey appeared to regain a normal use of the hand. The deficit, however,
could not be produced by a simple lesion to the cortical surface of area 6,
anterior to the primary motor hand area. Instead, it depended on a massive
removal of the lateral and medial surfaces of area 6 and the wedge of white
matter between them. Clearly the lesion disturbed some aspect of hand func-
tion thereby causing forced grasping, but the exact function that was disturbed
and its exact location in the brain were not clear.
Fulton (1935) suggested that the forced grasping produced by area 6 lesions
was part of a “premotor syndrome,” a loss of the highest levels of motor coor-
dination. In his own studies, Fulton reported that lesions to primary motor
cortex alone caused a temporary paralysis “most marked and enduring in the
distal joints.” Lesions to premotor cortex alone caused a temporary “disorga-
nization of the more integrated voluntary movements” and some effects on
visceral function. Lesions to both primary motor and premotor cortex, rather
than causing a temporary deficit, caused a “permanent paralysis of voluntary
movement.” The monkey became permanently locked in. In Fulton’s interpre-
tation of these lesion results, the primary motor cortex controlled simple move-
ment components whereas the premotor cortex controlled the coordination of
2. Early Experiments on Motor Cortex 31

complex motor acts. Although forming a natural motor hierarchy, the two
areas were able to function partially in parallel because after lesions of the pri-
mary motor area the premotor area was evidently able to take over some of
the lost function. Fulton speculated that there were two separate projection
systems from cortex to the spinal cord, one from primary motor cortex com-
monly called the pyramidal tract, and one from premotor cortex that Fulton
termed the “extrapyramidal tract.”

FOERSTER
The hypothesis of a premotor cortex was even more boldly stated in the writing
of Foerster (1936). During the 1920s and 1930s Foerster performed surgical
operations on the brains of humans to remove epileptic foci. The patients
were typically awake and under local anesthetic. Foerster was therefore able to
electrically stimulate the surface of the cortex and observe the effect on behavior.
Foerster (1936) stated without qualification, “The specific function of area 4 is
the isolated innervation of single muscle groups” (p. 152). In contrast, when
area 6 is electrically stimulated, “a complex mass movement of all parts of the
contralateral half of the body is obtained” (p. 148). The border between these
two areas in Foerster’s map (shown in Figure 2-7) is absolute.
In Foerster’s map only a type of reified myth is presented. Foerster was
describing what “everybody knows.” By claiming that complex and combined

Motor phenomen
a
S en
Fo eg

Head, eyes, trunk, to the s or

ot
L igh . Ch
Th do x

opposite side - Complex y

To ot
Ab ora ld.

Fo g
Th hou m

ph
e

synergic movement of
Le igh y
S Ar

Th Bell est

contralateral arm and leg Fo

H re Sh
C e no
D a ar o unl Disa m
D ig nd m F A uld oca gre
Di igi it 5 or r m e lize eab
en

r
B g t
Di it 4 D a ar H e d se le
a

Th git 3 Di igit nd m nsa

tion
um 2 Di gi 5
Ne b D git t 4 s
c i
Fac k Th git 2 3
Tonge um E
ue Neck b op yes
Jaw Face p to
Pala Mo
Lary te Ton uth sid osite
nx Ph gue e
a
Lary rym c
A nx usti s Com
Aco ination plex
uc hallu optic
hall o
nk t ex
yes, truCompl
cina
tion
s
D
d, e e- ent s
Heasite sid ovem metie
o m tre
oppynergic eral ex
s ralat
ont
of c

Figure 2-7 Somatotopic map of the human brain according to Foerster (from Vogt
and Vogt, 1926).
32 The Intelligent Movement Machine

movements were in the province of area 6, he was evidently forced to depict

the primary motor cortex in contrast as representing only simple and sepa-
rated movements. As a consequence, Foerster’s primary motor cortex was an
extraordinarily detailed one in which each separate fragment of the body, even
each finger, had its own cortical locus. Foerster represented one of the most
extreme forms of the view that the primary motor cortex is a muscle map of
the body.

DOUBTING PREMOTOR CORTEX: WALSHE AND PENFIELD

Fulton’s premotor cortex was not universally accepted. Walshe (1935) was
highly critical of Fulton’s lesion work, arguing that the “premotor syndrome”
was mainly a set of vague, general symptoms that occurred with brain lesions
almost everywhere, their very vagueness somehow convincing Fulton that they
were “high order”; and that Fulton had merely showed the existence of a gra-
dient in which posterior lesions in motor cortex had a larger effect, especially
on the fingers, than anterior lesions.
The validity of the premotor cortex was further questioned by Wilder Pen-
field and colleagues (Penfield and Boldrey, 1937; Penfield and Welch, 1951).
Penfield performed brain surgery on human patients under local anesthetic.
To map the brain before removing diseased tissue, he electrically stimulated
discrete points on the cortical surface and noted the evoked movements and
also the patients’ reports of induced sensation. His technique therefore was
the same as Foerster’s (with whom he sometimes worked), though his conclu-
sions were in some respects the opposite. In 1937, Penfield and Boldrey pub-
lished the most complete description yet of the motor map in the human
brain, based on 163 patients.
Penfield’s homunculus (Figure 2-8) has become stamped into the memory
of every neuroscience and medical student. His name is indelibly attached to
the motor map. This view of Penfield as the father of the motor map is obvi-
ously an exaggeration because he did not discover the motor map, nor was he
the first to demonstrate it in humans. He did however make at least six funda-
mental contributions to motor cortex physiology.
First, the size of the study and the meticulous manner in which it was reported
made the paper an instant classic.
Second, like many others before him, Penfield showed that movements and
sensations were partially overlapped in the cortex but that movements were
mainly evoked anterior to the central sulcus and sensations were mainly
evoked posterior to the central sulcus. In this way he helped to establish the
distinction between motor cortex and somatosensory cortex.
Third, Penfield argued against the view of the motor cortex as a roster of
perfectly separated body parts. Although in his map each body part is assigned
a separate location in the cortex, this drawing was not meant to be a literally
accurate depiction of the motor cortex. It was meant to summarize a general
trend. Penfield and Boldrey (1937) provided figure after figure of raw data
2. Early Experiments on Motor Cortex 33

Shoulder
Trunk

Elbow
Hip

Wrist
Kn

d
ee

n
Ha
An
kle

tle
To

M R ing
Lit
es

e
dl
ex

id
Ind mb
u
Th Neck
w
Bro ll
ba
eye
and
lid Face
Eye

V O C A L I Z AT I O N
Lips

Jaw
T
N

Sw ongu
AT I O

allo e
win
SA N

g
LIV
IO
AT

IC
ST
MA

Figure 2-8 The motor homonculus of the human brain from Penfield and Rasmussen
(1950). A coronal slice through the motor cortex is shown.

showing abundantly that the representations of adjacent body parts overlapped

and that the map of the body was of a statistical and blurred nature. This
aspect of Penfield’s work is often not appreciated. He himself dismissed the
novelty of the finding, pointing out that the overlap among body part repre-
sentations was already obvious in the earliest studies of motor cortex (Penfield
and Boldrey, 1937).
Fourth, Penfield rejected the division of motor cortex into a primary and a
premotor cortex. In his view it was not the case that simple, individual move-
ments were obtained in a distinct posterior area, or that complex, multiseg-
mental movements were obtained in a distinct anterior area. If any distinction
existed, it was a tendency for the representation of the fingers to be more con-
centrated in the posterior area and a representation of the trunk and shoulder
to be more concentrated in the anterior area, although even this distinction
34 The Intelligent Movement Machine

was blurred. Penfield may have gone too far in this direction of denying any
obvious distinction between posterior and anterior motor cortex. His position
appears to have been a reaction to the summary maps of Vogt and Vogt (1919)
and Foerster (1936), depicting a precise border between physiological zones,
with simple movements to one side of the border and complex movements
immediately to the other side. These maps appeared to be the result of imagi-
nation layered on top of a selective summary of the data. Penfield was correct
in pointing out the lack of any clear border between two distinct and uniform
areas. However, Penfield’s interpretation that no distinctions existed, and that
only one relatively homogenous motor field could be found in the lateral
motor cortex, may have been too extreme and certainly caused several decades
of controversy over the existence or nonexistence of the premotor cortex. To
Penfield, area 4 and area 6aα belonged together into one motor cortex; and
the more anterior field, area 6aβ, did not have motor functions at all. He sug-
gested that the complex, multisegmental movements that Vogt and Vogt
(1919) and Foerster (1936) had obtained from 6aβ were at least partly the
result of too much current inducing an epileptic seizure that then spread
indiscriminately into the motor cortex. The role of 6aβ (or PMDr as the
roughly corresponding area is now called) is still debated, though in contrast
to Penfield most researchers now believe that it plays at least some role in
motor control.
Fifth, Penfield and Welch (1951) proposed that although there was no pre-
motor cortex directly anterior to primary motor cortex, there was nonetheless
a second motor area located on the medial part of the hemisphere. Their
“supplementary motor area” (SMA) is shown in Figure 2-9. Their evidence for
two motor areas was based partly on what appeared to be two somatotopic
progressions of the body found in the monkey brain. In the lateral motor
map, progressing up the precentral gyrus, the representation moved in a
blurred fashion from the head, down the body, to the feet. In the medial motor
map, beginning at the posterior part of the precentral gyrus and progressing
forward, the representation moved in a blurred fashion from the feet, up the
body to the head. This medial map was different from the lateral map in that
it contained more overlap among adjacent body-part representations. The
extent of overlap was so great that although Penfield and Welch (1951)
reported the map in the monkey, they were unable to resolve the somatotopic
progression in the human SMA. The medial map could also be distinguished
from the lateral map by the types of movement evoked on stimulation. Stimu-
lation of the medial map tended to evoke multisegmental movements that
sometimes combined both sides of the body. Stimulation of the lateral map
seemed to evoke simpler movements of one side of the body. In Penfield’s
interpretation, therefore, at least some of the complex actions originally local-
ized to area 6aβ and attributed to a premotor function were more correctly
identified with the SMA.
A final important contribution of Penfield’s was the introduction of the term
homunculus to describe the motor map in the human, along with a drawing of
Exploring the Variety of Random
Documents with Different Content
 If a hastily prepared mush is required, perhaps nothing better
than the rolled oats can be employed, these requiring not more than
half an hour’s cooking, as they are already partially cooked in their
manufacture; but even these are improved by longer cooking in a
double boiler.
It is very important, when making any kind of mush, that the
water be boiling rapidly, and kept thus while stirring in the meal; for
unless the grain or meal is thoroughly scalded when stirred in, not
even prolonged cooking will take away the raw taste.

OATMEAL MUSH
To a quart of boiling water add a pinch of salt,
sprinkle in a cupful of oatmeal, and boil rapidly
for about ten minutes, or until it sets, stirring
frequently with a fork. Then place over the hot
water in the lower boiler and cook from one to
three hours. Just before serving, remove the Quart Measure
cover and stir lightly with a fork to allow the
steam to escape. This makes the mush more dry. Serve with
baked apples, cream, fresh fruit, or with the juice from stewed
fruit. Oatmeal is richer in nitrogen than any other grain, and
therefore very nutritious. But to be wholesome it must be well
cooked, and not served in a pasty, undone mass.

ROLLED OATS
This is much preferred by some, as it requires only a short
time to cook. Make as above, only using two cupfuls of the meal
to one quart of water. An ordinary saucepan does very well for
this, but the double boiler is better.
 ROLLED OATS AND SAGO MUSH
Wash and soak one-third cup of sago in a little cold water. Stir
one and one-half cups of rolled oats into one quart of salted,
boiling water. Cook for fifteen minutes, then stir in the sago, and
cook as much longer. Serve with cream, stewed fruit, or fruit
juice.

GRAHAM MUSH
Into three pints of rapidly boiling water, properly salted, stir
dry, one heaping pint of sifted Graham flour. Cook slowly for one
hour on the back of the range, stirring but little after the first
few minutes. Serve with milk or cream, and a very little sugar if
desired.

GRAHAM MUSH WITH DATES

Cook as above. Take a cupful of dates, cut in two, removing
the stones, and stir into the mush just before taking from the
fire. Serve with milk or cream. Steamed raisins or stewed figs
may be used instead of dates. Serve hot, or pour out into cups
or molds, first wet with cold water, and serve cold with cream.

BOILED RICE
Wash one cup of rice, and put to cook in four cups of boiling
water, slightly salted. Cook quite rapidly for the first fifteen
minutes, stirring a little occasionally to prevent sticking to the
pan. Then cover closely, and cook slowly on the back of the
range without stirring. When nearly done, add a cup of sweet
milk, cook until tender, and serve with milk, cream, or stewed
fruit. If the rice has been soaked overnight, put to cook in an
equal quantity of boiling water, or equal parts of milk and water,
and cook for about half an hour.

CREAM OF WHEAT
To four parts of boiling water previously salted, add one part
cream of wheat, sprinkling it in with the hand, and cook slowly
for about an hour. Serve hot with cream or stewed figs.

CORN-MEAL MUSH, NO. 1

Into three pints of boiling water, salted, sprinkle one pint of
corn-meal. Cook slowly for an hour, stirring occasionally. Serve
with plenty of milk or cream. Very good and nutritious,
especially for winter.

CORN-MEAL MUSH, NO. 2

Put to boil one quart of water, adding one teaspoonful of salt.
Mix smooth one tablespoonful of flour and two cupfuls each of
milk and corn-meal. Stir this gradually into the rapidly boiling
water; boil about half an hour, stirring frequently. Serve as soon
as done, with rich milk.

CORN-MEAL SQUARES
Take cold, left-over corn-meal mush, cut into rather thick
slices, and then into inch squares. Put the squares into a tureen,
and pour over them some hot milk or cream. Cover the dish, let
stand a few minutes, and serve.

BARLEY MUSH
 To each cupful of pearl barley, previously washed, add five
cups of boiling water, a teaspoonful of salt, and cook in a double
boiler for three or four hours. Serve with cream, lemon sauce,
or stewed fruit.

BOILED WHEAT
To one part of good, plump wheat add five parts of cold
water, a little salt, and cook slowly from four to six hours, or
until the grains burst open and are tender. If soaked overnight,
less time for boiling will be required. Add a little more water
while cooking if necessary, but avoid much stirring. Serve hot or
cold with milk, cream, fruit, or fruit juice. A very simple and
wholesome dish.

GLUTEN MUSH
Into three pints of rapidly boiling, salted water stir one pint of
gluten; cook in a double boiler for several hours.

HOMINY
Soak, then put to cook in enough boiling water to cover. Cook
gently for several hours, being careful not to stir after the grains
begin to soften. Add a little more water if needed. Season with
salt when done. A quantity may be cooked at a time, and
warmed up with a little cream or butter as needed.

CRACKED WHEAT
Cook the same as hominy and oatmeal, using three parts of
boiling water to one of cracked wheat. When done, turn into
cups or molds first wet with cold water. Nice served cold with
cream. Seedless raisins may be cooked with it.
 GRANULATED WHEAT
Use the same proportion and cook the same as cracked
wheat. Serve warm or cold with good sweet cream.

CORN-MEAL CUTLETS
Cut cold corn-meal mush into slices three inches long and one
inch wide; roll each piece in beaten egg, slightly salted, then in
grated bread crumbs; place on an oiled tin in the oven till nicely
browned. Other mushes may be treated likewise.

BROWNED RICE
Place a small quantity on shallow tins, and brown in the oven
till a golden yellow, stirring frequently so that it may brown
evenly; then steam for about an hour in a steamer over boiling
water or in a steam cooker, allowing two parts of hot water to
one part of rice. When done, it should be quite dry and mealy.
It may be eaten dry, or served with brown or lentil sauce, or
rich milk or cream.

BAKED MUSH
Cook any of the foregoing mushes as directed, and as soon as
done, turn into a pan, crock, or a round tin can, first wet with
cold water, or oiled, to prevent sticking. If brushed over the top
with oil, a crust will not form. When cold, cut into slices from
one half to three fourths of an inch thick, place on oiled tins,
and bake till a nice brown. A quart of cooked mush will make
about a dozen slices.
 “A meal—what is it? Just enough of food
To renovate and well refresh the frame,
So that with spirits lightened, and with strength renewed,
We turn with willingness to work again.”

The appetite is subject to education; therefore learn to love that which you
know to be good and wholesome.
The most expensive food is spoiled when served up burnt or tasteless; the
cheapest may be delicious with the proper seasoning.—Lantz.

T
makes a very nice breakfast dish, and is easily and quickly
oast
prepared. It can be made in a variety of ways which are both
simple and wholesome. When properly prepared, it furnishes
abundant nourishment, and is easily digested.
The proper foundation for all toasts is zwieback (pronounced
zwībäck), or twice-baked bread. This may be made from either fresh
or stale bread, the fresh making the more crisp and delicious for dry
eating. The bread should be light and of good quality. That which is
sour, heavy, and unfit to eat untoasted, should never be used for
toast.
 Toasts afford an excellent opportunity for using up left-over slices
of bread, and its use is therefore a matter of economy as well as of
securing variety in diet.

ZWIEBACK, OR DRY TOAST

Cut fresh or stale light bread, either white or brown, into
slices half an inch thick, place on tins, and bake slowly in a
moderate oven until browned evenly throughout. Care should
be taken not to scorch the bread. It should not be put into an
oven that is merely warm. It should be baked, not simply dried.
The common method of toasting merely the outside of the
bread by holding it over a fire is not the most wholesome way of
preparing toast. When properly made, it will be crisp
throughout. Zwieback may be prepared in quantity and kept on
hand for use. It furnishes a good article of diet, especially for
dyspeptics, eaten dry, or with milk or cream.

MILK TOAST
Scald one cupful of milk in double boiler, then add one
teaspoonful of cornstarch, mixed with a little cold water; stir
until it thickens. Cook about ten minutes, then add one
teaspoonful of butter, one-fourth teaspoonful of salt, and pour it
over six slices of zwieback, previously moistened with hot water
or milk.

TOAST WITH CREAM SAUCE

Prepare a cream sauce as directed on page 77. Moisten five
or six slices of zwieback by dipping them quickly into hot water
or milk, place them on a dish, and pour over the hot cream
sauce.

ASPARAGUS TOAST
Prepare asparagus by washing each stalk free from sand;
remove the tough portions, cut the stalks into small pieces, and
stew in a little hot, salted water; drain off the water as soon as
done, add a cup of milk, and season with a little butter and salt.
Cream may be used instead of the milk and butter. Moisten the
zwieback with hot milk, and place in a dish. Pour over the
stewed asparagus, and serve hot.

BERRY TOAST
Prepare zwieback as above. Take fresh or canned
strawberries, raspberries, mulberries, or other fruit, mash well
with a spoon, add sugar to sweeten, and serve as a dressing on
the slices of zwieback previously moistened.

EGG TOAST
Moisten slices of zwieback in hot milk or cream, season with a
sprinkle of salt, and serve hot with a poached egg on each slice.
For poached eggs see page 66.

BANANA TOAST
Moisten slices of zwieback in hot milk. Mash the bananas into
a pulp, or cut into thin slices, and place some on each slice of
toast.

FRUIT TOAST
 Take stewed apricots, peaches, or plums, rub through a
colander, heat to boiling, thicken with a little cornstarch,
sweeten to taste, and pour over the moistened zwieback.

CREAM TOAST
Moisten slices of zwieback in hot water, sprinkle with a little
salt, and dip over each slice a spoonful or two of nice, sweet,
cold cream.

BUTTER TOAST
Place each slice of zwieback on a small plate, pour over a
little hot water, and quickly drain off; add a sprinkle of salt, if
desired, spread lightly with butter and serve.

CRUSHED TOAST
Take fresh, but thoroughly toasted bread or crackers, or some
of each, grind closely in a coffee or hand mill, or crush with a
rolling-pin, and serve in small dishes with milk, cream, or fruit
juice. This may be served as a substitute for the health food
known as granola. Crushed toast is also a very serviceable
article for use in soups and puddings.

TOMATO TOAST
Moisten slices of zwieback in hot milk, and serve with a
dressing prepared by heating a pint of strained, stewed
tomatoes to boiling, and thickening with a tablespoonful of flour
or cornstarch rubbed smooth in a little cold water. Season with
salt and a little cream or butter, and pour over the toast.
 BEAN PASTE
Soak one cupful of white beans overnight in cold water; put
to cook in the morning in boiling water, and cook to a pulp, and
till the water is quite absorbed. Rub through a colander, then
add a tablespoonful of finely minced onion, one teaspoonful of
powdered sage, one saltspoonful of celery salt, the juice of one
lemon, two or three spoonfuls of tomato juice, if at hand, and
salt to taste. Simmer together for a short time, then use cold to
spread on toast or bread as a relish, or in the place of butter, or
for making sandwiches.
Variety.—Remember, as Home Note says, that “variety of diet
is important. Ill health often follows a monotonous sameness of
diet. Oatmeal, bread and butter, and marmalade, are all
excellent breakfast dishes of their kind, but when given every
morning, for years at a time, they become positively
nauseating.”
 A VOICE FROM THE CORN

“I was made to be eaten, not to be drank,

To be thrashed in a barn, not soaked in a tank;
I come as a blessing when put in a mill,
As a blight and a curse when run through a still;
Make me up into loaves, and your children are fed;
But made into drink, I will starve them instead.
In bread I’m a servant, the eater shall rule,
In drink I’m a master, the drinker a fool.
Then remember my warning; my strength I’ll employ,—
If eaten, to strengthen, if drunk, to destroy.”

The wandering Arab lives almost entirely upon bread, with a few dates as a
relish.
Behind the nutty loaf is the mill wheel; behind the mill is the wheat field; on the
wheat field rests the sunlight; above the sun is God.—James Russell Lowell.

B
readstands at the head of all foods. It has very properly been
termed “the staff of life.”
 Why this is so is because wheat, from which bread is mostly
made, contains more nearly than any other one article, all the
necessary food elements required to sustain the human system, and
these, too, in proper proportions, and so forms most nearly a perfect
food. From it the brain, bones, muscles, and nerves, all receive a
large amount of nourishment.
This being so, bread should enter largely into the daily bill of fare
of every family. It is hardly too much to say that no meal is complete
without it.
Where little bread is used, serious defects may frequently be
observed. For instance, in some of the islands of the Pacific Ocean,
where no wheat has been grown, and little could be obtained, the
inhabitants almost universally have poor teeth. The early decay of
the teeth so prevalent among the rising generation to-day, may
generally be attributed to four causes: (1) A lack of sufficient lime in
the water; (2) too free indulgence in sweets, such as rich cakes,
jams, and candies; (3) too large an amount of flesh foods; and (4)
an insufficient supply of good, simple, wholesome bread, especially
whole wheat bread.
Home-made bread, when properly prepared, is generally to be
preferred to bakers’ bread. Chemicals and adulterations, as well as a
lack of cleanliness and proper care in preparation, not infrequently
characterize the latter, and thus give rise to serious stomach
disorders. Moreover, bakers’ bread is not always obtainable, and is
always necessarily more expensive than that which is home-made.
The baker can not afford to work for nothing. For these reasons,
every woman, and especially every wife and mother, ought to know
how to make good bread. The temptation to patronize the bake
shop should not outweigh the interests of the health of the family,
and the duty to practise economy.
The essentials to good bread-making are three:—

1. Good flour.
 2. Good yeast.
3. Proper attention.

When either of these is lacking, good results can not be obtained.

Poor flour will not produce good bread; good flour and poor yeast
will not make good bread; and good flour and good yeast with
improper attention will not insure good bread. All three are essential.
The first thing to consider in the making of bread is the flour.
Good flour will generally be found to have a creamy white tint. That
which is of a bluish white is seldom the best. Good flour will fall light
and elastic from the hand. Flour that retains the imprint of the
fingers when squeezed, and falls in a damp, clammy mass, should
be avoided.
The second essential is good yeast. One may have ever so good
flour and yet make poor bread, if the flour is used in conjunction
with poor yeast. Good yeast has a fresh, pungent odor, and is light
and foamy; while poor yeast has a sour odor, and a dull, watery
appearance.
The third essential is proper attention. In winter, bread sponge
should be made at night if it is desired to have the bread baked in
the early part of the day. The flour used in making the sponge
should first be warmed, and the sponge covered with several
thicknesses of cloth, and set in a warm place till morning.
In hot weather set the sponge early in the morning, and the bread
can be baked by noon. Both the sponge and dough are best kept in
an earthen crock or jar, as they are less quickly affected by drafts of
air.
As soon as the sponge has risen to be light and puffy, it should
receive attention immediately, if desired to have the bread white and
sweet. If allowed to reach the point of running over, or falling in the
center, it has stood too long. For this reason sponge set at night
should be mixed late in the evening, and attended to as early in the
morning as possible.
 In using very active yeast, it will not be necessary to set a sponge.
Mix the ingredients into a good bread dough at the first mixing,
beating the batter well while stirring in the flour. The more
thoroughly the batter is beaten, the less kneading the dough will
require. Set the bread in this way in the morning, and it can be
baked by noon.
A few mealy potatoes, cooked and mashed, added to the sponge,
makes the bread sweeter and keeps it fresh longer. Milk used in
connection with yeast should first be scalded and cooled to
lukewarm.
Too much flour should not be used in mixing, as it will make the
bread hard and tough; but enough should be used to make the
dough firm and elastic. Turn the dough out on the molding-board
and knead it, not with the tips of the fingers, but with the whole
hands, from the sides into the center, turning frequently, that all
portions may be thoroughly worked. When the dough is smooth and
elastic, with no dry flour left on its surface, form into a smooth ball,
and place back in the crock, which should be washed clean, dried
and oiled, to prevent the dough from sticking. Observe how full it
makes the crock; cover up warmly, and when it has doubled its bulk,
form gently into loaves, handling the dough as little as possible, and
place in the pans for the last rising. When the loaves are risen to
twice their size, place in a moderately hot oven to bake. The oven
should be hot when the bread is put in. By no means have the
bread, when ready to bake, wait for the oven to be heated, as it
may then become too light, run over in the oven, and possibly be
sour.
When nearly ready to bake, test the oven by putting in it a piece
of writing-paper; if it turns dark brown in six minutes, the oven is of
about the proper heat. If bread bakes too fast, a crust is formed on
the outside of the loaf which prevents the inside from becoming hot
enough to dry thoroughly, and the result is that the inside of the loaf
is too moist, while the outside is baked hard. Bread should not
brown much under fifteen or twenty minutes after being placed in
the oven. If it rises much after being put in the oven, the heat is not
sufficient. Bread should be turned around in the oven if it does not
rise or brown evenly.
Medium-sized loaves should be baked from fifty to sixty minutes;
small French loaves about thirty-five minutes. Bread is done when it
shrinks from the pan, and can be handled without burning the
fingers.
When taken from the oven, the loaves should be turned out of the
pans, placed on their sides, so that the crust will not soften by the
steam, and covered with a thin cloth. When cold, keep in a covered
stone jar or a tin box, which should be kept free from crumbs and
musty pieces of bread, and scalded and dried thoroughly every few
days.
As to their healthfulness, the most wholesome breads are
unleavened breads, or those made without either yeast, baking-
powder, soda, or cream of tartar, such as gems, rolls, and crackers.
Next come those made with good yeast; then those with baking-
powder, if comparatively pure; and lastly those made with soda and
sour milk, or soda and cream of tartar. Baking-powder is preferable
to soda. The latter should seldom if ever be used, as it is injurious to
the health, being an active dyspepsia-producing article.

WHITE BREAD
Scald a quart of new or unskimmed milk, let cool
to lukewarm, then stir in a dissolved yeast cake,
two teaspoonfuls of salt, and enough sifted flour to
make a thin batter. Cover, and set aside till light,
then work in flour until a dough of the proper
Flour Sieve consistency for bread is formed. Knead until it is
smooth and elastic, and does not stick to the hands
or board. Place in a clean, oiled crock, and when light, form into
four loaves; let rise again and bake. Equal parts of milk and
water may be used if desired.

MOTHER’S BREAD
In the evening boil three small potatoes, or save them out
when cooking, and mash them with a fork in a gallon crock. Put
in about three cupfuls of flour, two tablespoonfuls each of salt
and sugar, then pour in enough boiling water to make a good
batter. Beat until smooth. Soak one cake of compressed yeast or
yeast foam in one-half cup of lukewarm water, and when the
batter is just warm stir in the yeast and beat until quite foamy.
Set in a warm place overnight. The first thing in the morning dip
about two quarts of flour in a pan, make a cavity in the center,
and pour in the sponge and about a pint of warm water. Stir all
together into a thin batter, and set in a warm place till after
breakfast; then knead until it does not stick to the board, put it
in a three-gallon crock, well oiled to prevent the dough from
sticking; cover with a tin lid to keep a crust from forming over
the top, then with several thicknesses of cloth, and set in a
warm place until it rises up full. Then mold into loaves, place in
pans, let rise again, and bake in a moderate oven for about an
hour, or until the loaves shrink from the sides of the pans and
do not burn the fingers when removing from the pans. Turn the
bread out of the pans, and cover with a thin cloth. This will
make six loaves. If the loaves are brushed over with cold water
just before being placed in the oven the crust will be more crisp.

Baking Pan

GRAHAM BREAD, NO. 1

 Take two tablespoonfuls of good liquid yeast, two cups of
sweet milk, previously scalded and cooled to lukewarm, one
teaspoonful of salt, and two cupfuls of white flour; beat
together thoroughly, and set to rise. When very light, add three
heaping cupfuls of sifted Graham flour, or sufficient to make a
soft dough. Knead for a half-hour, then place in a pan slightly
buttered, cover warmly, and set to rise. When light, form into
loaves, let rise again, and bake.

GRAHAM BREAD, NO. 2

Make a sponge as for white bread. When light, add the stiffly
beaten white of one egg, one tablespoonful each of sugar and
melted butter, and enough sifted Graham flour to make a soft
dough. Knead lightly, place back in oiled crock till light, then
make into loaves, let rise, and bake. Graham bread should not
be mixed as stiff as white bread, or it will be too solid. Two
tablespoonfuls of molasses may be used for sweetening instead
of sugar, if preferred.

GRAHAM FRUIT BREAD

Make the same as Graham bread, and when ready to form
into loaves, add a cupful of raisins or dried currants, washed
and dried, and dusted with flour.

WHOLE WHEAT BREAD

Make a sponge as for white bread. If desired a light color, use
one fourth white flour instead of all whole wheat flour. Knead
well, keeping the dough soft, then set in a warm place to rise.
When light, form into loaves, let rise again, and bake. This
bread rises slower than white bread.
 BOSTON BROWN BREAD
Scald one pint of corn-meal with a pint of boiling
water; let cool till lukewarm, then stir in one
dissolved yeast cake, or one-half cup of sweet,
lively yeast, three tablespoonfuls of molasses, one
Pint Measure
teaspoonful of salt, and about three cupfuls of rye
meal. Beat well, put in oiled pan, steam four or
five hours, then place in the oven for half an hour
to form a crust.

PARKER HOUSE ROLLS

Take two cupfuls of lukewarm milk, previously scalded, three
tablespoonfuls of melted butter, or vegetable oil, one well-
beaten egg, three tablespoonfuls of sugar, a pinch of salt, and
one cake of yeast dissolved in a little of the milk; mix all
together, then add enough flour to make a good batter. Let rise
until light, knead, using sufficient flour; let rise again till very
light, roll out to one-half inch in thickness, cut into round or oval
shapes with a cutter, fold one third back over the top, and place
in a pan to rise. When very light, bake in a moderate oven.
Brush over with beaten yolk of egg, mixed with two spoonfuls of
cold water just before taking from the oven. Braided or plaited
rolls may be made by cutting the rolled dough into strips six
inches long and one inch wide, pinching the ends of each three
strips together, and then braiding.

CORN-MEAL BREAD
Stir one-half cup of corn-meal into two cupfuls of boiling
water; when well cooked, remove from the fire and add two
cupfuls of cold water; stir well together; then add one
teaspoonful of salt, one cake of yeast dissolved in a little warm
water, two tablespoonfuls of sugar or molasses, and enough
white flour to make a good dough. Knead well, and set to rise;
when light, form into three loaves, let rise again, and bake for
nearly an hour.

SALT-RISING BREAD
Take a small pitcher and put into it a half pint of warm water,
a teaspoonful each of salt and sugar, then stir in flour enough to
make a medium-thick batter. Set the pitcher in a kettle of warm
water to rise. It should be kept warm all the time, not hot, for if
it is scalded, it will never rise. When light, stir in a pint of warm
milk or water and enough warm flour to make a soft dough.
Knead it, form into a loaf, place in the pan, set to rise in a warm
place, and bake as soon as light.

RAISED BISCUITS
Make from dough prepared for white bread. When the dough
is ready to form into loaves, divide it into small, equal portions,
shape into smooth, round biscuits, place closely in a shallow
baking pan, and let rise till considerably lighter than bread;
brush lightly with milk, and bake in a rather quick oven.

GEMS
General Directions
Beating in an abundance of cold air is very essential in the
making of good gems, as it is this that makes them light. Cold
air is preferable to warm air, as it expands more when heating.
Gems are also better when baked in iron pans than in tin, as
the iron retains the heat better, and bakes the gems more
evenly. The irons should be heated and oiled
before the batter is dropped into them.
Having the oven hot from the first is also Gem Irons
essential, as a crust will then be formed
immediately, and the air which has been beaten into the batter
will thus be prevented from escaping. They should be placed in
the oven so as to bake on the top first, and afterward on the
bottom. These points should be carefully observed. Gems are
best served hot. They should be broken open, and never cut
with a knife, as this makes them heavy.

GRAHAM GEMS, NO. 1

Place the gem irons in the oven or on the range to heat. Mix
salted Graham flour with cold milk or water to a batter thick
enough to drop, beating vigorously for ten minutes to beat in
the air. Butter the gem irons, and fill each cup nearly full of the
batter. Put in a hot oven, and bake until done.

GRAHAM GEMS, NO. 2

Beat separately the yolk and white of an egg. Add to the

beaten yolk two cupfuls of sweet, rich milk, one-half teaspoonful
of salt, and stir well together; then sift in one and one-half cups
of Graham flour, and a scant cup of white flour, beating
vigorously meanwhile. Continue to beat until the mixture is light
and foamy throughout, and full of air bubbles; then stir in gently
the stiffly beaten white of the egg. Have the gem irons
thoroughly heated, slightly butter them, drop in the batter with
a spoon, and bake in a quick oven.

OATMEAL GEMS
 Beat separately the yolk and white of an egg. To the beaten
yolk add a cupful of well-cooked oatmeal mush, and a half cup
of milk or thin cream. Beat together thoroughly. Continue to
beat while adding a cupful of white flour and a pinch of salt,
then fold in lightly the stiffly beaten white of the egg. Have the
gem irons heated hot, slightly butter, drop in the batter, filling
the little cups nearly full, and bake in a quick oven until a light
brown.

CORN-MEAL GEMS

Stir well together one and one-half cupfuls of milk, and the
yolks of two eggs previously beaten. To this add two cupfuls of
corn-meal, one-half teaspoonful of salt, and one cupful of white
flour. Beat thoroughly, then stir in lightly the whites of the eggs
previously beaten to a stiff froth, and bake as above.

GRANULATED WHEAT GEMS

Mix together one cupful each of cold water and milk, and one-
half teaspoonful of salt. Then add gradually two and one-half
cupfuls of fine granulated wheat, beating continuously. Beat
vigorously for ten minutes, then drop by spoonfuls into
thoroughly heated, buttered gem irons, beating the batter
briskly several times while dipping it in. Bake at once in a very
hot oven.

RICE CAKES
Moisten one cup of well-cooked rice with two tablespoonfuls
of cream or rich milk; add one tablespoonful of sugar, and mix
in enough flour to make it hold together. Form into cakes one-
third of an inch thick, and bake in a hot oven. When done, split
open, and serve with maple or lemon sirup. To make lemon
sirup, see page 40.
 BREAKFAST ROLLS
To three slightly heaping cups of sifted Graham flour add a
little salt, and one cup of milk or thin cream; cream is better.
Stir the milk or cream into the flour, mixing it well with the flour
as fast as poured in. Knead thoroughly, then divide the dough
into three portions, and with the hands roll each portion over
and over on the molding-board until a long roll from an inch to
an inch and a half in thickness is formed. Cut into two- or three-
inch lengths, and bake at once in a hot oven, in a baking pan
dusted with flour, or better, on a perforated piece of sheet-iron
made for the purpose, placing the rolls a little distance apart.
Bake until a light brown. When done, do not place one on top of
another.
Flour kneaded into cold Graham flour, oatmeal, or corn-meal
mush makes very good breakfast rolls.

STICKS
Make the same as breakfast rolls, only rolling the dough to
about the size of the little finger, and cutting into three- or four-
inch lengths.

FRENCH ROLLS
Make a sponge at night of one-half cake of dry or one-half
cup of good liquid yeast, the beaten white of one egg, two
tablespoonfuls of melted butter, two tablespoonfuls of sugar, a
little salt, and three cups of warm milk or water, and flour
sufficient to make a soft dough. In the morning knead well and
let rise again. When light, roll out the dough to about three
fourths of an inch in thickness; cut into about four-inch squares
with a sharp knife, butter the edges, and roll each corner up
and over to the center; place on buttered tins, allow the rolls to
become very light, and bake in a moderately hot oven. The
sponge for this can be set in the morning if the yeast is very
quick.

TO GLAZE ROLLS
When ready to bake, brush the rolls or biscuit lightly with
milk; or, when nearly baked, brush with the yolk of an egg to
which has been added two spoonfuls of cold water and half a
teaspoonful of sugar. Return to oven till done.

MARYLAND OR BEATEN BISCUIT

Mix five cupfuls of white flour, one-half cupful of vegetable oil
or butter, and one teaspoonful of salt to a very stiff dough with
one cupful of cold water. Knead for twenty minutes, using no
more flour for the molding-board; then beat hard with a
wooden mallet or hammer for twenty minutes longer, until the
dough is flat and of even thickness throughout; sprinkle over a
little flour, fold half of the dough back evenly over the other half,
and beat quickly around the edges, to keep in the air. Continue
beating until the dough is brittle, and will snap if a piece is
broken off quickly. Pinch off into pieces the size of a small
walnut, work smooth, flatten on top with the thumb, prick with
a fork, place on perforated tins a little distance apart, and bake
in a moderate oven for nearly an hour, or until dry and brittle
throughout.

WHOLE WHEAT CRISPS

Take one cupful of rich cream, two tablespoonfuls of sugar, a
pinch of salt, two cupfuls, or enough to make a stiff dough, of
fine granulated, whole wheat flour. Beat well, and knead for
fifteen minutes, first with a spoon, until the batter becomes too
thick, and then with the hands. Roll out as thin as wafers, cut
into shapes with a biscuit cutter, and bake on floured tins in a
very hot oven.

GRAHAM WAFERS
Stir together one cupful each of sifted Graham
flour and white flour, one tablespoonful each of
butter and sugar, and a saltspoonful of salt; then
mix with enough cold water to make a stiff dough.
Cake Cutter Roll out very thin, cut into small squares, or with a
cake cutter, and bake on tins in a quick oven.

FRUIT BISCUIT
Make a dough with one cupful of cold, sweet cream or rich
milk, three cupfuls of sifted Graham or white flour, and a little
salt. Knead thoroughly, and divide into two portions. Roll each
quite thin, then spread one with currants, stoned dates, figs, or
seedless raisins, chopped fine, and place the other one on top;
press down with the rolling-pin, cut into oblong squares with a
knife, and bake.

CRESCENTS
Make a dough, using the recipe for White Bread. When ready
to form into loaves, work into it two tablespoonfuls each of
butter and sugar; roll out into a sheet half an inch thick, cut into
six-inch squares, then divide diagonally, forming triangles; brush
each lightly with water, and roll up, beginning at the longest
side; place on oiled pans, turning the ends toward each other in
the form of a crescent. When very light, brush with milk, and
bake in a quick oven for about twenty minutes.
 RUSKS
Make a sponge at night with one cupful of sugar, one cupful
of scalded milk, cooled to lukewarm, one-half cupful of butter,
two eggs, one cake of dry or one-half cup of good liquid yeast,
and sufficient flour to make a drop batter. Set in a warm place
to rise. In the morning knead well, and when risen again, mold
into the form of biscuits, place a little distance apart on buttered
tins, and brush over with the beaten white of an egg
sweetened; let stand until light, and bake.

PLAIN BUNS
Beat together one-fourth cup of lively yeast, one cup of sweet
milk, previously scalded and cooled to lukewarm, one-half
teaspoonful of salt, two cups of warm flour, and set in a warm
place to rise. When very light, work into the dough one-half cup
of sugar, and two tablespoonfuls of butter. Knead well for ten
minutes, using enough flour to make a soft dough. Shape into
the form of biscuits a little larger than an egg; place on tins
slightly buttered, and set in a warm place to rise. When very
light, bake in a moderately hot oven. The tops may be brushed
over with the sweetened beaten white of an egg while baking,
or sprinkled with moist sugar when taken from the oven.

FRUIT BUNS
Make the same as plain buns, adding one-half cup of raisins
or currants just before kneading and forming into buns.

RICE WAFFLES
Set a sponge at night with two cupfuls of sweet milk, scalded
and cooled to lukewarm, one tablespoonful of butter, a pinch of
 salt, two-thirds of a cupful of boiled rice, three
cupfuls of flour, and one-fourth cup of liquid yeast.
Beat the batter hard for five or six minutes, and
set in a warm place to rise. In the morning add
two well-beaten eggs, and stir well together. Bake
on a hot, buttered waffle iron. If this is not at
Waffle Iron hand, have the gem irons well heated, slightly
butter to prevent sticking, and drop in the batter.
Place in a hot oven so the top will bake first, and bake to a rich
brown color. Very nice for breakfast.

PUFFS
To two cups of milk add a little salt and the yolks of two eggs
well beaten; then sift in, a little at a time, and beating
meanwhile, three small cups of flour. Beat until light, then stir in
gently the stiffly beaten whites of the eggs, and bake in hot
gem irons.

FRUIT LOAF, NO. 1

Take enough good bread dough for one loaf, add one cupful
of brown sugar, two tablespoonfuls of butter, and one cupful of
raisins, previously washed and dried. Knead well and let rise;
then knead again, and place in a bread pan, let rise until light,
and bake in a moderate oven.

FRUIT LOAF, NO. 2

Make a sponge of one and one-half cups of warm milk or
water, one-half cup of good yeast, the beaten white of one egg,
one tablespoonful each of butter and sugar, a little salt, and
flour sufficient to make a soft dough. Let rise till light; then
knead well and let rise again. When light, roll out to about one
inch in thickness, spread over with chopped dates, or raisins, or
currants which have been previously washed and dried; roll up
and form into a loaf, let rise, and bake.

COFFEE CAKES
Take two cupfuls of bread dough (made with milk) when
ready for the pans; put into a deep dish and work in four
tablespoonfuls of cocoanut or vegetable oil or butter, four
tablespoonfuls of sugar, the stiffly beaten white of one egg, and
enough flour to make a fairly stiff dough. Knead well, and roll
out into a long strip about nine inches in width, three feet in
length, and one fourth of an inch thick; spread over this four or
five tablespoonfuls of oil or melted butter, omitting about two
inches at the farther end; beginning at end nearest, roll up like
jelly roll; cut into slices an inch thick; place a little distance apart
on tins sprinkled with sugar; set in a warm place, and when
very light, brush over with oil; sprinkle with a little sugar, and
bake. If desired, ground cinnamon or grated nutmeg may be
sprinkled over the dough before rolling it up.

FLANNEL CAKES
Heat three cupfuls of milk to boiling; put into a crock one
cupful of corn-meal and two tablespoonfuls of butter, then pour
in the scalding milk; beat well, allow to cool to lukewarm, then
stir in one tablespoonful of sugar, two of flour, one teaspoonful
of salt, and one-half yeast cake dissolved in one-third cup warm
water; beat well, and set to rise overnight. Bake on a hot
griddle.

CORN-MEAL BATTER CAKES

 To two cups of cold corn-meal mush, add one
cup of sifted flour, and a pinch of salt; beat well
the yolks of two eggs, to which add two-thirds cup Griddle
of milk, and stir into the mush; beat thoroughly
until light and smooth, adding a little more milk if
necessary, to make the batter of proper consistency. Then
gently stir in the whites of the eggs beaten to a stiff froth, and
bake in small cakes on both sides on a griddle, slightly buttered,
or better still on a soapstone griddle, in which case use no oil
nor butter on it. Serve hot.

BUCKWHEAT PANCAKES
In the evening take two quarts of warm water, add one-fourth
cup of good yeast, a teaspoonful of salt, and buckwheat flour
enough to make a good batter. If desired, a cupful of corn-meal
or a few spoonfuls of white flour may be used instead of all
buckwheat. Beat well and set to rise. In the morning thin the
batter with a little warm water, if necessary, and bake on a hot
griddle. If cakes are desired for several mornings, the batter
may be kept going by leaving at least a cupful after each
baking, and adding the necessary warm water and buckwheat
flour each evening as at first.

LENTIL FRITTERS
To a pint of lentil soup (left-over soup will do), add the well-
beaten yolks of two eggs, and sift in enough flour, a little at a
time, beating thoroughly, to make a good batter. Then add the
stiffly beaten whites of the eggs, drop by spoonfuls on a hot
buttered griddle, and brown on both sides.

CORN FRITTERS