Midterm 2 Practice Answers

Splooey
March 2024

1 Introduction
This midterm practice is composed of 3 parts:

1. Review - 10 true or false questions. Covers content from before Midterm 2, but will be useful
for Midterm 2
2. Elementary Questions - 10 true or false questions and 3 long computation questions. Content
that you should know for Midterm 2

3. Advanced Questions - 5 questions of different levels of difficulty. For anyone who is interested,
will not appear on Midterm 2 and will not be helpful. In fact, they will give you a headache
and make your life worse.
Starred questions are harder than non-starred questions. Currently, there are only solutions for the
Review and Elementary Questions.
Good luck on Midterm 2 guys!

1
2 Review
2.1 True or False
1. All column spaces are subspaces (True)
2. All null spaces are subspaces (True)
*3. The 1 × 3 matrix [2, 3, 4] is a subspace (False, a matrix is not a subspace)
*4. 2x + 3y + 4z = 0 is a subspace (True, it is the null space for [2, 3, 4])
 
a
b
 
5. R5 is a subspace of R7 (False, vectors in R5 are something like  7
 c  whereas vectors in R are

d
e
 
a
b
 
c
 
 d )
something like  
e
 
f 
g
6. All subspaces are spanned by a basis (True)
7. Not all subspaces are vector spaces (False)
8. If V and U are subspaces of Rn , V + U is a subspace of Rn (True)
9. If V is a subspace of R5 with dim2 and U is a subspace of R5 with dim3, V + U spans R5
(False, what if V lies on U ?)
10. If V is a vector space with dim3, V is R3 (False, V can be a subspace in R4 with dim3,
remember, subspaces are also vector spaces)

3 Elementary Questions
3.1 True or False
1. Let there be the linear transformation T : V → W , V and W have the same dimensions. Then,
T is onto if and only if T is one-to-one (True)

n o T : V → W ,nT o
*2. Let there be the linear transformation (x) = Ax where A is a matrix. T is
⃗ ⃗
one-to-one if and only if NulA = 0 (True, NulA = 0 means every column in the REF for
A has a pivot, so T is one-to-one)
3. Determinants can only be found for square matrices (True)
4. Square matrices are invertible if and only if all rows and all columns are linearly independent
(True, they imply that the REF for a square matrix would have pivots for every row and every
column respectively)

2
 5. f (x) = x3 − x is onto. (True, for every y there exists at least one x such that y = x3 − x; but
also read the footnote 1 )
6. f (x) = x3 − x is one-to-one. (False, when y = 0, x = −1, 0, 1 can all create y; but also read
the footnote 1)

*7. Let there be the linear transformation T : V → W , V and W have the same dimensions. Sup-
pose T is onto. Then, if {x⃗1 , x⃗n1 } isolinearly independent, {T (x⃗2 ), T (x⃗2 )} is also linearly
n inde-
o
pendent. (True, Assume that ⃗a, ⃗b is a linearly independent set. Suppose that T (⃗a), T (⃗b)
is not linearly independent. Then, T (⃗a) is a linear combination of T (⃗b). Then, T (⃗a) = cT (⃗b).
By properties of linear transformation, T (c⃗a) = cT (⃗a, which
n would
o imply that ⃗a is a linear
combination of ⃗b, which cannot be true since we assumed ⃗a, ⃗b to be linearly independent.
Hence, (7.) must be true.)
 
*8. Let c1 , c2 .c3 .c4 be 4 columns of a 4 × 4 matrix A such that A = c1 c2 c3 c4 . Sup-
pose detA = 6. Then, det (2c1 + 2c4 ) (c2 ) (c3) (2c1 + 3c4 ) = 12 (True. By mul- 
tilinear
 in columns property for determinants,
  det (2c 1 + 2c4 ) (c2 ) (c3 ) (2c1 + 3c4 ) =

det (2c1 ) (c2 ) (c3 ) (2c1 + 3c4 ) + det (2c4 ) (c2 ) (c3 ) (2c1 + 3c4) and  
det (2c1 ) (c2 ) (c3 ) (2c1 +  3c4 ) +det (2c4 ) (c2 ) (c3 ) (2c1 + 3c4 ) = det (2c1 ) (c2 )  (c3 ) (2c1 ) +
det (2c1 ) (c2 ) (c3 ) (3c4 ) +det (2c4 ) (c2 ) (c3 ) (2c1 ) +det (2c4 ) (c2 ) (c3 ) (3c4 ) .
We know that matrices with columns  that
 are ”linearly dependent”  have a determinant of zero, 
so det
 (2c 1 ) (c2 ) (c3 ) (2c
1 ) +det (2c
 1 ) (c2 ) (c3 ) (3c 4 ) +det
 (2c
 4 ) (c2 ) (c3 ) (2c1 ) +
det (2c4 ) (c2 ) (c3 ) (3c4 ) = 0+det (2c1 ) (c2 ) (c3 ) (3c4 ) +det (2c4 ) (c2 ) (c3 ) (2c1 ) +
0 and swapping rows multiply the determinant by -1. So, the determinant is 2(3)(6)−2(2)(6) =
12. Also, read footnote. 2 )

9. The transpose of all upper triangular matrices are lower triangular matrices. (True)
 
*10. Suppose an n × n matrix L = λ1 λ2 . . . n λon where each λ is a column where all of the
entries except one row is 0. Then, NulL = ⃗0 if and only if all λi ̸= 0 and there does not
exist a λi such that λi and λn
j contain
o the non-zero entry on the same row. (True. We have
⃗
shown earlier that if NulL = 0 , L represents a one-to-one transformation. We defined L as
an n × n matrix so L also represents an onto transformation. This means that none of the
lambdas can be zero as this would produce a column of zero and make the transformation
L represents neither onto nor one-to-one. This fulfills one of the two conditions. The other
condition just means that none of the vectors have λ in the same row, which would mean every
row has a λ and so does every vector λ, hence every column in L since they are column vectors
for L)

1 The original question actually said x3 − x is onto, in which, this is just an expression and not a function since I

did not set it to f (x), so technically you would be right saying that it does not make sense to call it onto or one-to-one
2
 Originally it should have been  false since the original question forgot to include det in front of
(2c1 + 2c4 ) (c2 ) (c3 ) (2c1 + 3c4 ) = 12. Since a matrix that is not not 1 × 1 cannot be a number, it is
automatically false

3
3.2 Computation
3.2.1 Matrices, invertibility and basis changing.
a. Find a matrix A for a linear transformation T : R2 → R2 that takes the vectors e⃗1 and e⃗2 and
 into ⟨3, 1⟩ and ⟨4, 5⟩.
turns them
3 4
Answer:
1 5

b. Find a matrix B for a linear transformation T : R2 → R2 that takes ⟨3, 1⟩ and ⟨4, 5⟩ turns
them into e⃗1 and e⃗2 . Is Ainvertible? What is the inverse of A if there exists one?
5/11 −4/11
Answer: , yes, A−1 = B
−1/11 3/11

c. Suppose A takes the e⃗1 and e⃗2 and turns them into ⟨a, 1⟩ and ⟨4, b⟩, a and b are real numbers.
Write an expression in terms of a and b for all a and b such
 thatA is non-invertible.
a 4 a 4
Answer: This is the same as finding such that det = 0. We know the determi-
1 b 1 b
 
a 4
nant for two by two matrices can be calculated as follows: det = ab − 4(1). So, for all
1 b
ab = 4, the determinant for the matrix is 0 and it is invertible. Therefore, the answer is any
ab = 4.
2 2
*d. Suppose A is a matrix for n T : oR → V , and B is a matrix for T : V → R . Let V be a
⃗ ⃗ ⃗ ⃗
vector space spanned by b1 , b2 where b1 and b2 are linearly independent and are standard
coordinate vectors for V . Find b⃗1 and ⃗
 b2 interms of standard coordinates for R .
2

1 0
Answer: We want to transform and from basis B to the standard, so we can use our
0 1
matrix B to do so.
    
5/11 −4/11 1 5/11
= and so on. In fact, we can just spot them by looking at each
−1/11 3/11 0 −1/11    
5/11 −4/11
column for our B matrix. So, b⃗1 = and b⃗2 = in standard coordinates.
−1/11 3/11

3.2.2 Determinants
 
1 3 2 5
2 3 5 1
a. Find the determinant of   by row reduction. Is this matrix invertible?
2 3 −4 8
−1 2 −3 1  
1 3 2 5
2 3 5 1
Answer: Do a bunch of row replacements to  , we get the following:
2 3 −4 8
−1 2 −3 1
 
1 3 2 5
0
 −3 1 −9
0 −3 −8 −2
0 5 −1 6

4
 Do some more row replacements and we get
 
1 3 2 5
0 −3 1 −9
 
0 0 −9 7
0 0 −8/3 −9

One more row replacement and we get

 
1 3 2 5
0 −3 1 −9 
 
0 0 −9 7 
0 0 0 −229/27

The matrix is now upper triangular, and notice we have not done any row multiplications nor
swapping, so this matrix and our original have the same determinant. Now, just multiply the
entries on the diagonal and we get (1)(−3)(−9)(−229/27) = −229 which is our answer. The
determinant is non-zero so it is invertible.
 
a 0 1 b −3
3 0 0 4 0
 
b. Find the determinant of −7 3 9 4 c  (Hint: Your answer is a number).

2 0 0 0 0
d 0 2 3 1
 
0 1 b −3
0 0 4 0 
Answer: Cofactor expansion along the fourth row, we get detA = 2det  3 9 4 c 


0 2 3 1
Now expand along the first column  (you can  also expand along the second row here by the
1 b −3
way) and we get detA = 2(3)det 0 4 0 . And then expand along the second row and get
  2 3 1
1 −3
detA = 2(3)(4)det .
2 1
 
1 b −3
detA = 2(3)det 0 4 0  = 2(3)(1 + 6) = 42
2 3 1

So 42 is our answer.
c. Suppose detA = 6, A is 5 × 5. Find det(2A−1 )
Answer: detA = 1/detA−1 ; detA−1 = 1/6. Remember that det(cA) = cn detA. A is 5 × 5, so
det(2A−1 ) = 25 (1/6) = 16/3
d. Suppose a quadrilateral in R2 with vertices at (4, 5), (−3, 5), (9, −5) and (−3, −2). Find the
area of this quadrilateral.
Answer: Split the quadrilateral into two halves. The first half is the triangle between the
vertices (4, 5), (−3, 5) and (9, −5), the second is the triangle between (−3, 5), (9, −5) and
(−3, −2). For the first one, start from (4, 5) and create two vectors connecting to (−3, 5)
and (9, −5). The vectors are ⟨−7, 0⟩ and ⟨5, −10⟩. For the second one, start from (−3, 5)

5
 and create two vectors connecting to (9, −5)and (−3, −2).
 The vectors are ⟨12, −10⟩ and the
−7 5
second is ⟨0, −7⟩. Find the determinant of and divide it by 2, you get 35. Find
  0 −10
−12 0
the determinant of and divide it by 2, you get 42. Add the numbers up and you
−10 −7
get 77, which is the area of the quadrilateral.
*e. Find the volume of an ellipsoid given by the equation ( 2x−y 2 y+2z 2 z−x 2
3 ) + ( 4 ) + ( 2 ) = 13
3 3
Answer: Let T : R → R be a linear transformation, T ⃗x = A⃗x. Let R denote a region, and
S denote a sphere. Then, the volume of S = T (R) = |detA|volume of R. Let R be the region
3

x2 + y 2 + z 2 = 13, which gives a sphere with radius (13). Then, the volume of S is 4(13)
2
p
3 π.
Let x, y and z be our directions in R3 . Then, we can do the following transformation based
on how x, y, z transforms in ( 2x−y 2 y+2z 2 z−x 2
3 ) + ( 4 ) + ( 2 ) = 13:
 
2/3 0 −1/2
A = −1/3 1/4 0 
0 1/2 1/2
And so if we get the absolute value of this determinant, we can find our area. The determinant
3
4(13) 2 1 3
(calculations skipped) is 1/6, so 3 π = 6 volume of R. Therefore, 8(13) 2 π is the final
answer.

3.2.3 Eigenvalues, eigenvectors and eigenspaces

*a. Find a 5 × 5 matrix such that there are only the eigenvalues 3 and 2, the eigenspace for λ = 3
is 1 dimensional and the eigenspace for λ = 2is 2 dimensional.

3 ∗ ∗ 0 0
0 3 ∗ 0 0
 
Answer: One guaranteed solution would be  0 0 3 0 0 where ∗ can be any non-zero

0 0 0 2 0
0 0 0 0 2
number and can be different. There are also other solutions, but this one is one of the more
straightforward ones.
 
1 2 0 1
−1 3 − 1 0 2
b. Write out, but do not factor, the characteristic polynomial of 
8
.
3 2 0
0 0 2 1
Answer: First write the matrix A − λI, which is the following:
 
1−λ 2 0 1
 −1 2−λ 0 2 
 
 8 3 2−λ 0 
0 0 2 1−λ
Now, we need to set det(A − λI) = 0, where det(A − λI) is expressed as the characteristic
polynomial of A.
To find det(A − λI), first, do cofactor expansion along the third column.
   
1−λ 2 1 1−λ 2 1
det(A − λI) = (2 − λ)det  −1 2−λ 2  − 2det  −1 2 − λ 2
0 0 1−λ 8 3 0

6
 After some more cofactor expansions, we get
     
1−λ 2 2 1 1−λ 1
det(A − λI) = (2 − λ)(1 − λ)det − 2(8)det + 2(3)det
−1 2−λ 2−λ 2 −1 2
And taking the determinants of the 2 × 2 matrices, we get the following:

det(A − λI) = (2 − λ)(1 − λ)((1 − λ)(2 − λ) + 2) − 2(8)(2 + λ) + 2(3)(3 − 2λ)

Expanding all the terms, we get the following:

det(A − λI) = λ4 − 6λ3 + 15λ2 − 46λ − 6

And that is our answer.

 
3 2 3
c. Find all eigenvalues and eigenvectors for 0 2 1, list the algebraic and geometric multi-
0 0 2
plicities for each eigenvalue.
Answer: Read footnote 3 . First, observe that the matrix is upper triangular, so the entries on
the diagonals are the eigenvalues. We see that we have eigenvalue 3 with algebraic multiplicity
1 and eigenvalue 2 with algebraic multiplicity 2.
 which is Nul(A − λI). For eigenvalue 3, the eigenspace is the null
Now, find the eigenspace,
0 2 3
space of 0 −1 1 . By observation, we see that the eigenspace is one dimensional and
0 0  −1
   
 1  1
spanned by 0 so any non-zero scalar multiple of 0 is an eigenvector for 3, eigen-
0 0
 
value
 3 has  a geometric multiplicity of 1. For eigenvalue 2, the eigenspace is the null space
1 2 3
of 0 0 1. By observation (or by calculation if you like), the eigenspace is spanned by
 0 0 0  
 −2  −2
 1  . So, any non-zero scalar multiple of  1  is an eigenvector for eigenvalue 2, and
0 0
 
eigenvalue 2 has a geometric multiplicity of 1.  
1
So, the overall answer is the eigenvalues are 3 for any non-zero scalar multiple of 0 and 2
  0
−2
for any non-zero scalar multiple of  1 . The eigenvalue 3 has an algebraic multiplicity of 1
0
and a geometric multiplicity of 1, whereas the eigenvalue 2 has an algebraic multiplicity of 2
and a geometric multiplicity of 1.
     
 3 1   3 
*d. Suppose the vector spaces V and U in R3 spanned by 2 , 2 and −2 respectively.
0 1 w
   
Find a w such that the intersection of V and U can never be an eigenvector for all 3×3 matrices.
3 The question was changed due to it being nearly impossible to solve by hand.

7
 Answer: The zero vector can never be an eigenvector, so we want
 to find a w such that the
3
only intersection between V and U is the zero vector, aka −2 is not in the span of V . So,
w
we can put our vectors into an augmented matrix as follows:
 
3 1 3
2 2 −2
0 1 w

After some reductions, we get the following:

 
1 1 −1
0 1 −3 
0 0 w+3

When w is -3, the

 augmented matrix does not have a pivot in the last column. Therefore, for
3
all w ̸= −3, −2 is not in the span of V and therefore the intersection of V and U can never
w
be an eigenvector for all 3 × 3 matrices.
e. Are all eigenspaces subspaces? Why or why not?
Answer: All eigenspaces are subspaces since they are the null space for the matrix A − λI.
They also include the zero vector even though the zero vector is not an eigenvector.

4 Advanced Questions
4.1 Jordan normal form
The Jordan normal form is a given by a matrix J such that for all matrices A for a transformation
T : Cn → Cn , there exists an invertible matrix P such that A = P JP −1 . A matrix A in Jordan
normal form looks as follows:
 
J1

 J2 

J(A) = 
 J3 

 .. 
 . 
Ji

Where all areas that are not part of a Ji are filled with zeros and each Ji is a Jordan block in the
following form:  
λ 1 0 ... 0
 .. .. 
0
 λ 1 . .
Jr =  ... ..
 
 0 λ . 0 
. .. ..
 ..

. . 1
0 ... ... 0 λ

8
where all λ in a given Jordan block are the same, there are n λs on the diagonal, λ is an eigenvalue
for a matrix A, Jr is r × r, the diagonal above λ is filled with 1. If λ has an algebraic multiplicity r
for the matrix A, λ has a geometric multiplicity of 1.
*a. Let a Jordan block be an r × r matrix called Jr . Then, show that Jr has a nullity of 0 for all
λ ̸= 0.
*b. Show that for all matrices A where the algebraic multiplicity of the eigenvalues overall are
equal to the geometric multiplicity of the eigenvalues, all Jordan blocks are 1 × 1
*c. Show that for all matrices A where the algebraic multiplicity of the eigenvalues overall are
equal to the geometric multiplicity of the eigenvalues, the Jordan normal form of A is the
following:  
λ1

 λ2 

J(A) = 
 λ 3



 . .. 

λn
Where there are precisely n eigenvalues if algebraic multiplicity is counted, J(A) is a diagonal
matrix.

9
4.2 The Fibonacci Sequence
(From MATH 223) Let the Fibonacci sequence be denoted f1 , f2 , f3 . . . fn where f1 = 1, f2 = 1, fi =
fi−1 + fi−2 for all i is an integer. Let the following be a matrix equation:
    
1 1 fi f
= i+1
1 0 fi−1 fi
   
fn f
**a. Write, but do not solve for an equation for computing using the vector 2 .
fn−1 f1
 
1 1
b. Find a matrix D which contains the eigenvalues of on the diagonal and a matrix P ,
1 0
the eigenvalue in column 1 of D has an eigenvector which is the column vector of column 1 of
P.
*c. Find and simplify an equation for fn using the information above in the form of abn − acn .
Simplify this equation further by taking the limit of acn as n approaches infinity.

4.3 Integers modulo n

Let n be a fixed positive integer. We will define a relation on integers such that a ≡ b mod n if and
only if a − b is divisible by n. We write ā as the set of all as for a ± kn, k ∈ Z and call elements of ā
equivalence classes. Suppose that ā + b̄ = a + b and āb̄ = ab.

**a. Assume V is a vector space and W is a subspace and let there be the equivalence relation
v ≡ v ′ if v ′ − v ∈ W . Show that the set v + W = {v + w; w ∈ W } is a subspace (in particular,
it is called an affine subspace). (From MATH 223)
**b. Solve the following system of equations in Z/8Z:

[5]x + [3]y = [2]
[7]x − [2]y = [3]

(From MATH 322)

**c. Let n ∈ Z, n > 1 and let a ∈ Z, 1 ≤ a ≤ n. Show that if a and n have no common factors
other than 1 (in other words, they are relatively prime), there exists an integer c such that
ac ≡ 1 mod n

4.4 Groups
A group is an ordered pair (G, ∗) with a binary operation operation ∗ on G which fulfills the following
for all x, y, z ∈ G:

x ∗ x̃ = i, x ∗ i = x (Contains an inverse and identity element)

(x ∗ y) ∗ z = x ∗ (y ∗ z) (Associativity)
A group is called abelian or commutative if and only if x ∗ y = y ∗ x for all x, y ∈ G. A subgroup is
a group within a group.

10
**a. Show that (N, ·), N = {All n × n invertible matrices}, · is the operation for matrix multipli-
cation, is a group.
Z ×
**b. Show that for n ∈ Z+ , the set ( nZ ) of equivalence classes ā with multiplicative inverses mod
n is an abelian group under multiplication.

**c. Let GL2 (R) = {A|A is 2 × 2 with entries from R} and let there be a binary operation between
any two elements of GL2 which is matrix multiplication (GL2 (F ) where  the
 entires of A are 
a b
from any field is called the general linear group of degree 2). Let G = |a, b, c ∈ R and ̸= 0 .
0 c
Show that G is a subgroup of GL2 (R) and that GL2 (R) is closed under matrix multiplication
and inverses.
**d. Let (G, ∗) and (H, ⋄) be groups. A map ϕ : G → H such that ϕ(x ∗ y) = ϕ(x) ⋄ ϕ(y) for all
x, y ∈ G is called a homomorphism. If ϕ is a homomorphism and bijective, ϕ : G → H is an
isomorphism written G ∼ = H. Show that G is abelian if and only if H is abelian.

4.5 Fields
A field is a triple (F, +, ·) which fulfills the following for all x, y, z ∈ F

(x + y) + z = x + (y + z); x + y = y + x; x + x′ = 0; x + 0 = x

(x · y) · z = x · (y · z); x · y = y · x; x · x̃ = 1, x ̸= 0; 1x = x
x(y + z) = xy + xz; (x + y)z = xz + yz
**a. Show that there exists only one identity element for both addition and multiplication and show
that there exists a unique additive and multiplicative inverse for all x ̸= 0.
n o
**b. Find the smallest possible set S such that S is a field, S ̸= ∅, ⃗0 . Justify your answer.

**c. Let X be a set and let a power set P(X) = {A|A ⊂ X} be defined as the collection of all
1x∈A
subsets A ⊂ X. Associate an indicator function 1A (x) = to a subset A ⊂ X. Show
0x∈/A
that the map A → 1A is a bijection between P(X) and a vector space FX
2 (From MATH 223)

**d. Show that if n is not prime, Z/nZ is not a field.

11