FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING

BENJAMIN HACKL, CLEMENS HEUBERGER, SARA KROPF, AND HELMUT PRODINGER

arXiv:1704.01095v1 [math.CO] 4 Apr 2017

ABSTRACT. Rooted plane trees are reduced by four different operations on the fringe. The
number of surviving nodes after reducing the tree repeatedly for a fixed number of times is
asymptotically analyzed. The four different operations include cutting all or only the leftmost
leaves or maximal paths. This generalizes the concept of pruning a tree.
The results include exact expressions and asymptotic expansions for the expected value and
the variance as well as central limit theorems.

1. INTRODUCTION
Plane trees are among the most interesting elementary combinatorial objects; they appear
in the literature under many different names such as ordered trees, planar trees, planted
plane trees, etc. They have been analyzed under various aspects, especially due to their rel-
evance in Computer Science. Two particularly well-known quantities are the height, since it
is equivalent to the stack size needed to explore binary (expression) trees, and the pruning
number (pruning index), since it is equivalent to the register function (Horton-Strahler num-
ber) of binary trees. Several results for the height of plane trees can be found in [3, 9, 23],
for the register function, we refer to [4, 11, 19], and for results on the connection between
the register function and the pruning number to [4, 27].
Reducing (cutting-down) trees has also been a popular research theme during the last
decades [17, 21, 22]: according to a certain probabilistic model, a given tree is reduced until
a certain condition is satisfied (usually, the root is isolated).
In the present paper, the point of view is slightly different, as we reduce a tree in a com-
pletely deterministic fashion at the leaves until the tree has no more edges. All these reduc-
tions take place on the fringe, meaning that only (a subset of) leaves (and some adjacent
structures) are removed. We consider four different models:
– In one round, all leaves together with the corresponding edges are removed (see
Section 2).
2010 Mathematics Subject Classification. 05A16; 05C05 05A15 05A19 60C05.
Key words and phrases. Plane trees, pruning, tree reductions, central limit theorem, Narayana polynomials.
B. Hackl and C. Heuberger are supported by the Austrian Science Fund (FWF): P 24644-N26 and by the Karl
Popper Kolleg “Modeling-Simulation-Optimization” funded by the Alpen-Adria-Universität Klagenfurt and by
the Carinthian Economic Promotion Fund (KWF).
H. Prodinger is supported by an incentive grant of the National Research Foundation of South Africa. Part
of this author’s work was done while he visited Academia Sinica. He thanks the Institute of Statistical Science
for its hospitality.
This is the full version of the extended abstract [14].
1
2 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

– In one round, all maximal paths (linear graphs), with the leaves on one end, are
removed (see Section 3). This process is called pruning.
– A leaf is called an old leaf if it is the leftmost sibling of its parents. This concept was
introduced in [2]. In one round, only old leaves are removed (see Section 4).
– The last model deals with pruning old paths. There might be several interesting
models related to this; the one we have chosen here is that in one round maximal
paths are removed, under the condition that each of their nodes is the leftmost child
of their parent node (see Section 5).
The four tree reductions are illustrated in Figure 1. We describe these reductions more
formally in the corresponding sections.

Leaves Paths Old leaves Old paths

FIGURE 1. Removal of (old) leaves / paths

The first model is clearly related to the height of the plane tree, and the second one to
the Horton-Strahler number via the pruning index [27, 24]. While there are no surprises
about the number of rounds that the process takes here, we are interested in how the fringe
develops. The number of leaves and nodes altogether in the remaining tree after a fixed
number of reduction rounds is the main parameter analyzed in this paper.
For the sake of simplicity, we will use the same notation for each of the following reduction
analyses. In case we need to compare objects from two different sections, we will distinguish
them by adding appropriate superscripts.
The random variable X n,r models the tree size after reducing a plane tree of size n (that is
chosen uniformly at random among all trees with n nodes) r-times iteratively according to
one of our four reductions. If a tree does not “survive” r rounds of reductions, we consider
the size of the resulting tree to be 0. In particular, for r = 0, the given plane tree is not
changed and X n,0 = n.
As we will see, a key aspect of the analysis of X n,r is the translation of the algorithmic
description of the reduction into an operator Φ that acts on the corresponding generating
functions.
In Section 2, the reduction cutting away all leaves from the tree is discussed. Section 2.1
contains all necessary auxiliary concepts required in order to study the r-fold application of
this reduction. In Section 2.2, we determine the operator Φ acting on the corresponding
generating function explicitly and prove some direct consequences. Then, in Section 2.3 we
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 3

carry out the analysis of the behavior of X n,r by computing explicit expressions and asymptotic
expansions for the factorial moments of X n,r as well as a central limit theorem.
Section 3 is devoted to the study of the reduction that cuts away all paths. As we will see in
Section 3.1, we can actually obtain all results regarding the behavior of X n,r as consequences
of the corresponding results in Section 2. In Section 3.2, we analyze the asymptotic behavior
of the expected number of paths required to construct a plane tree of size n, i.e. the number
of paths we can cut away until the tree cannot be reduced any further.
Sections 4 and 5 are devoted to the analysis of reductions removing only leftmost leaves
and leftmost paths from the tree, respectively. In particular, in Section 5.3, we study the total
number of old paths that can be removed from a tree until it cannot be reduced any further.
On a general note, the computationally heavy parts of this paper have been carried out
with the open-source computer mathematics system SageMath [5], and the corresponding
worksheets are available for download. In particular, there are the following files:
• treereductions.ipynb for most of the asymptotic computations in Sections 2, 3,
and 4,
• old_paths.ipynb for most of the asymptotic computations in Section 5,
• factorial_moments_leaves.ipynb for computation of the factorial moments in
Theorem 1,
• factorial_moments_old_paths.ipynb for computation of the factorial moments
in Theorem 6.
Additionally, in order to run these computations yourself, you also need to download the
following two utility files:
• identities_common.py,
• conditional_substitution.py.
All these files including some instructions on how to use them can be found at https://
benjamin-hackl.at/publications/treereductions/.

2. CUTTING LEAVES
2.1. Preliminaries. In this part of the paper we investigate the effect of the tree reduction
that cuts away all leaves from a given tree. However, before we can do so, we require some
auxiliary concepts, which we discuss in this section. Most importantly, we need a generating
function counting plane trees with respect to their number of inner nodes and leaves, which is
intimately linked to Narayana numbers. The generating function presented in the following
proposition is actually well-known (see, e.g. [12, Example III.13]).

Proposition 2.1. The generating function T (z, t) which enumerates plane trees with respect to
their internal nodes (marked by the variable z) and leaves (marked by t) is given explicitly by
p
1 − (z − t) − 1 − 2(z + t) + (z − t)2
T (z, t) = . (1)
2
4 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

Proof. This can be obtained directly from the symbolic equation describing the combinatorial
class of plane trees T , which is illustrated in Figure 2. In particular, ƒ and represent leaves
and internal nodes, respectively.

= +
P
T ƒ
n≥1

T T T ··· T

FIGURE 2. Symbolic equation for plane trees

The symbolic equation translates into the functional equation

zT (z, t)
T (z, t) = t + ,
1 − T (z, t)
which yields (1) after solving for T (z, t) and choosing the appropriate branch. 

In the context of plane trees, the so-called Narayana numbers count the number of trees
with a given size and a given number of leaves (cf. [8]). As these numbers will appear
throughout the entire paper, we introduce them formally and investigate some properties
within the following statements.
Definition. The Narayana numbers are defined as
1
 ‹ ‹
n n
Nn,k =
n k−1 k
for 1 ≤ n and 1 ≤ k ≤ n, and N0,0 = 1. All other indices give Nn,k = 0. Combinatorially, for
n ≥ 1 the Narayana number Nn,k corresponds to the number of plane trees with n edges (i.e.
n + 1 nodes) and k leaves. The Narayana polynomials are defined as
n
X
Nn (x) = Nn,k x k−1
k=1

for n ≥ 1 and N0 (x) = 1, and the associated Narayana polynomials are defined as
Ñn (x) = x · Nn (x)
for n ≥ 0. Note that
1 2n
 ‹
Nn (1) = Ñn (1) = Cn =
n+1 n
is the nth Catalan number.
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 5
p
Remark. The generating function 1z T (z, z) = 1− 2z1−4z enumerates Catalan numbers, see [7,
Theorem 3.2], and the generating function T (z, tz) enumerates Narayana numbers
n−1
XX X
T (z, tz) = z t + Nn−1,k z n t k = z n Ñn−1 (t). (2)
n≥2 k=1 n≥1

We will frequently use this relation in the form

X €tŠ
T (z, t) = z n Ñn−1 . (3)
n≥1
z

Furthermore, it is easily checked that T (z, tz) satisfies the ordinary differential equation
∂
(1 − 2(t + 1)z + (1 − t)2 z 2 ) T (z, tz) − ((1 − t)2 z − t − 1)T (z, tz) = t(1 + z − tz).
∂z
Extracting the coefficient of z n+2 then yields the recurrence relation
(n + 3)Ñn+2 (t) − (2n + 3)(t + 1)Ñn+1 (t) + n(t − 1)2 Ñn (t) = 0 (4)
for n ≥ 0.
The following proposition gives another useful property of associated Narayana polynomi-
als.
Proposition 2.2. Let n ≥ 0, then we have the relation
€1Š
n+1
t Ñn = (1 − t)[n = 0] + Ñn (t). (5)
t
Proof. This relation follows from extracting the coefficient of z n+1 from the identity T (tz, z) =
T (z, tz) + (1 − t)z with the help of (3).
While it is straightforward to prove that the identity is valid by means of algebraic manip-
ulation, we also give a combinatorial proof.
From a combinatorial point of view, both generating functions T (tz, z) and T (z, tz) enu-
merate plane trees where z marks the tree size, the only difference is that the variable t
enumerates inner nodes in T (tz, z) and leaves in T (z, tz). We want to show that for trees of
size n ≥ 2, these two classes are equal, resulting in T (tz, z) − z = T (z, tz) − tz.
To construct an appropriate bijection between the class of trees of size n with k leaves
and the class of trees of size n with k inner nodes we need to have a closer look at the well-
known rotation correspondence [12, I.5.3], which is a bijection between plane trees of size n
and binary trees with n − 1 inner nodes. In fact, the leaves in the binary tree are strongly
related to the leaves and inner nodes of the original tree:
– Left leaves in the binary tree are only attached to those nodes whose companions in
the plane tree have no children, i.e., to those who correspond to leaves in the plane
tree.
6 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

– Right leaves, on the other hand, are attached to nodes whose companion nodes in
the plane tree have no sibling right of them. This means that for every node with
children, i.e., for every inner node, there is precisely one rightmost child and thus
precisely one right leaf in the binary tree.
The bijection between the two tree classes can now be described as follows: given some tree
of size n and k leaves, apply the rotation correspondence in order to obtain a binary tree.
Then mirror the binary tree by swapping all left and right children. Transform this mirrored
tree back by means of the inverse rotation correspondence, and the result is a plane tree of
size n and k inner nodes as mirroring the binary tree swapped the number of left and right
leaves in the tree. This proves the proposition. 
Derivatives of the associated Narayana polynomials defined above will occur within the
analysis of a reduction model later, which is why we compute some special values in the
following proposition.
Proposition 2.3. Evaluating the rth derivative of the associated Narayana polynomials at 1,
i.e. Ñn(r) (1), gives the number of trees with n + 1 nodes where precisely r leaves are selected and
labeled from 1 to r. In particular, for n ≥ 1 we have
1 2n 2n − 2
 ‹  ‹
Ñn (1) =
0
, Ñn (1) = (n − 1)
00
.
2 n n−1
Proof. The combinatorial interpretation follows immediately by rewriting
n
X
Ñn(r) (1) = Nn,k k r ,
k=1

where we used the notion k r = k(k − 1) · · · (k − r + 1) for the falling factorial. Explicit
values can be obtained by differentiating (2) r-times with respect to t, then setting t = 1 and
extracting the coefficient of z n+1 . 
2n
Remark. By the combinatorial interpretation of Proposition 2.3 we find that Ñn0 (1) = 12 n

enumerates the number of

1 2n

leaves, summed over all trees with n + 1 nodes. At the same
2n

time,
as there are Cn = n+1 n such trees, the total number of nodes in these trees is n . This
implies that exactly half of all nodes in all trees of given size are leaves!
In fact, this interpretation also motivates a second, purely combinatorial proof of the ex-
plicit value of Ñn0 (1): the bijection correspondence maps trees of size n + 1 to binary trees
with n inner nodes. In the proof of Proposition 2.2 we already observed that the number
of left leaves in the binary tree obtained from the rotation correspondence is equal to the
number of leaves in the plane tree.
As binary trees with n inner nodes have n + 1 leaves, and as there are Cn binary trees
with
2n
n inner nodes, the total number of leaves in all binary trees with n inner nodes is n . By
symmetry,
there have to be equally many left

leaves as right leaves—which proves that there
1 2n 1 2n
are 2 n left leaves, and thus Ñn (1) = 2 n .
0
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 7

In addition to the polynomials related to the Narayana numbers, there is another well-
known sequence of polynomials that will occur throughout this paper.

Definition. The Fibonacci polynomials are recursively defined by

F r (z) = F r−1 (z) + zF r−2 (z)

for r ≥ 2 and F0 (z) = 0, F1 (z) = 1.

For many identities involving Fibonacci numbers, there is an analogous statement for Fi-
bonacci polynomials. The identity presented in the following proposition will be used re-
peatedly throughout this paper.

Proposition 2.4 (d’Ocagne’s Identity). Let s, r ∈ Z≥0 where s ≥ r. Then we have

F r+1 (z)Fs (z) − F r (z)Fs+1 (z) = (−z) r Fs−r (z). (6)

F r+1 (z) F r (z)

 ‹
Proof. The left-hand side of (6) can be expressed as the determinant of . At
Fs+1 (z) Fs (z)
the same time, for r, s ≥ 1 we can write

F r+1 (z) F r (z) F r (z) F r−1 (z) 1 1

 ‹  ‹ ‹
= .
Fs+1 (z) Fs (z) Fs (z) Fs−1 (z) z 0

Combining these two observations yields

‹r
F r+1 (z) F r (z)
 ‹  ‹ 
1 0 1 1
F r+1 (z)Fs (z) − F r (z)Fs+1 (z) = det = det det ,
Fs+1 (z) Fs (z) Fs+1−r (z) Fs−r (z) z 0

which proves the statement. 

Observe that setting s = r + 1 in (6) yields the identity

F r+1 (z)2 − F r (z)F r+2 (z) = (−z) r , (7)

which we will make heavy use of later on.

An important tool in the context of plane trees is the substitution z = u/(1 + u)2 , which
allows us to write some expressions in a manageable form. It is easy to check that with this
substitution, we can write Fibonacci polynomials as

1 − ur
F r (−z) = . (8)
(1 − u)(1 + u) r−1

The fact that this substitution also works for Fibonacci polynomials is not that surprising,
as z F r (−z)/F r+1 (−z) is the generating function of plane trees with height ≤ r (see [3]).
8 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

7→ 7→ 7→

FIGURE 3. Illustration of the “cutting leaves”-operator ρ

2.2. Leaf-Reduction and the Expansion Operator. The reduction ρ : T \ { } → T we

want to investigate now can be explained very easily. For any tree τ ∈ T \ { } we obtain the
reduced tree ρ(τ) simply by removing all leaves from τ. Repeated application of ρ to a tree
is illustrated in Figure 3.
It is easy to see that this operator is certainly not injective: there are many trees that
reduce to the same tree. However, it is also easy to see that ρ is surjective, as we can always
construct an expanded tree that reduces to any given tree τ by attaching leaves to all leaves
of τ.
In fact, the operator ρ −1 mapping trees τ ∈ T to the set of preimages is easier to handle
from a combinatorial point of view. This is because we can model the expansion of trees in
the language of generating functions.
Proposition 2.5. Let F ⊆ T be a family of plane trees with bivariate generating function
f (z, t), where z marks inner nodes and t marks leaves. Then the generating function of ρ −1 (F ),
the family of trees whose reduction is in F , is given by
€ z zt Š
Φ( f (z, t)) = (1 − t) f , . (9)
(1 − t)2 (1 − t)2
Proof. It is obvious from a combinatorial point of view that the operator Φ has to be linear.
Thus we only have to determine how a tree represented by an arbitrary monomial z n t k , i.e.
a tree τ with n inner nodes and k leaves, is expanded.
In order to obtain all possible tree expansions from τ, we perform the following operations:
first, all leaves of τ are expanded by appending a nonempty sequence of leaves to each of
them. Then, every inner node of τ is expanded by appending (possibly empty) sequences of
leaves between two of its children as well as before the first and after the last one.
In terms of generating functions, expanding the leaves of τ corresponds to replacing t by
z t/(1 − t). Expanding the inner vertices is a bit more involved: by considering that every
inner node has precisely one more available position to attach new leaves than it has children
we find that there are 2n + k − 1 available positions overall within τ. Therefore we find
€ z t Šk 1
Φ(z n t k ) = z n ,
1 − t (1 − t)2n+k−1
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 9

which, as Φ is linear, immediately proves (9). 

Corollary 2.6. The generating function for plane trees T (z, t) satisfies the functional equation
T (z, t) = t + Φ(T (z, t)). (10)
Proof. This follows directly from the fact that ρ : T \ { } → T is surjective, i.e. ρ −1 (T ) =
T \ { }. 
Corollary 2.7. The Narayana numbers satisfy the identity
k 
2n + k − ` − 2
X ‹
Nn+k−1,k = Nn−1,`
`=1
k−`
for n ≥ 2, k ≥ 1.
Proof. The result follows from extracting the coefficient of z n t k from both sides of (10). 
Remark. Note that in [1] there is a very short proof based on Dyck paths for this identity, and
actually the argumentation there is strongly related to our tree reduction here: by the well-
known glove bijection, it is easy to see that cutting away all leaves of a plane tree translates
into removing all peaks within the corresponding Dyck path.
We are now interested in determining a multivariate generating function enumerating
plane trees with respect to the tree size as well as the size of the tree after applying the
tree reduction ρ a fixed number of times.
Proposition 2.8. Let r ∈ N0 . The trivariate generating function G r (z, vI , vL ) = G rL (z, vI , vL )
enumerating plane trees whose leaves can be cut at least r-times, where z marks the tree size,
and vI and vL mark the number of inner nodes and leaves of the r-fold cut tree, respectively, is
given by
1 − u r+2 € u(1 − u r+1 )2 u r+1 (1 − u)2 Š
G r (z, vI , vL ) = Φ r (T (zvI , t vL ))| t=z = T vI , vL .
(1 − u r+1 )(1 + u) (1 − u r+2 )2 (1 − u r+2 )2
(11)
Proof. First, observe that formally, we can obtain the generating function enumerating plane
trees that can be reduced at least r-times with respect to their size by considering Φ r (T (z, t))| t=z .
If we additionally track some size parameter like the number of inner nodes or the number
of leaves before the expansion by marking their size with vI and vL , then we obtain a gen-
erating function for plane trees that can be reduced at least r-times where vI and vL mark
inner nodes and leaves in the original tree and z marks the size of the expanded tree. From
a different point of view, z marks the size of the original tree and vI and vL mark the number
of inner nodes and leaves of the r-fold reduced tree, meaning that we have
G r (z, vI , vL ) = Φ r (T (zvI , t vL ))| t=z ,
which proves the first equation in (11).
10 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

As Φ is linear, we are mainly interested in finding a representation for Φ r (z n t k )| t=z . To do

so, we consider the strongly related operator
€ z zt Š
Ψ( f (z, t)) := f , .
(1 − t)2 (1 − t)2
It is easy to prove by induction that iterative application of Φ can be expressed in terms of Ψ
via
r−1
Y
Φ ( f (z, t)) = Ψ ( f (z, t))
r r
(1 − Ψ j (t)),
j=0

which means that we can concentrate on the investigation of the linear operator Ψ. Note
that Ψ is also multiplicative, meaning that Ψ r (z n t k ) = Ψ r (z)n Ψ r (t)k .
Again by induction, it is easy to show that the recurrences
zΨ r (t) z
Ψ r+1 (t) = Q r and Ψ r+1 (z) = Q r
j=0
(1 − Ψ j (t))2 j=0
(1 − Ψ j (t))2
hold for r ≥ 0. Now define f r := Ψ r (t)| t=z and g r := Ψ r (z)| t=z . We prove by induction that
these quantities can be represented by means of Fibonacci polynomials as
z r+1 zF r+1 (−z)2
fr = and gr =
F r+2 (−z)2 F r+2 (−z)2
for r ≥ 0, where the recurrence relations from above, the identity (7) as well as the relation
r−1
Y F r+2 (−z)
(1 − f j ) =
j=0
F r+1 (−z)

for r ≥ 0 play integral parts in the proof.

With these explicit representations, we find
r−1
Y z n+k(r+1) F r+1 (−z)2n−1
Φ (z t )| t=z
r n k
= Ψ (z t )| t=z
r n k
(1 − f j ) = . (12)
j=0
F r+2 (−z)2n+2k−1

Then, using (8) and rewriting the right-hand side of (12) in terms of u, where z = u/(1 + u)2 ,
yields
1 − u r+2 € u(1 − u r+1 )2 Šn € u r+1 (1 − u)2 Šk
Φ r (z n t k )| t=z = .
(1 − u r+1 )(1 + u) (1 − u r+2 )2 (1 − u r+2 )2
By linearity, we are allowed to apply Φ r to every summand in the power series expansion of
f (z, t) separately—which proves the statement. 
The generating function G r (z, v, v) tells us how many nodes (marked by v) are still in the
tree after r reductions. For the sake of brevity we set G r (z, v) := G r (z, v, v). It is completely
described in terms of the function T (z, t), although in a non-trivial way. Results about mo-
ments and the limiting distribution can be extracted from this explicit form.
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 11

With the help of the mathematics software system SageMath [5], the generating function
G r (z, v) can be expanded. For small values of r, the first few summands are
G1 (z, v) = vz 2 + (v 2 + v)z 3 + (v 3 + 3v 2 + v)z 4 + (v 4 + 6v 3 + 6v 2 + v)z 5 + O(vz 6 ),
G2 (z, v) = vz 3 + (v 2 + 3v)z 4 + (v 3 + 5v 2 + 7v)z 5 + (v 4 + 7v 3 + 18v 2 + 15v)z 6 + O(vz 7 ),
G3 (z, v) = vz 4 + (v 2 + 5v)z 5 + (v 3 + 7v 2 + 18v)z 6 + (v 4 + 9v 3 + 33v 2 + 57v)z 7 + O(vz 8 ).
As announced in the introduction, we investigate the behavior of the random variable
X n,r = X n,r
L
that models the number of nodes which are left after reducing a random tree τ
with n nodes r-times. In case the r-fold application of ρ to τ is not defined, we consider the
resulting tree size to be 0, i.e., the random variable X n,r = 0 for these trees. Note that the tree
τ is chosen uniformly at random among all trees of size n. With the help of the generating
function G r (z, v) we are able to express the probability generating function of X n,r as
an,r + [z n ]G r (z, v)
Ev X n,r
= (13)
Cn−1
where an,r is the number of trees of size n which are empty after reducing r-times. We have
an,r = Cn−1 − [z n ]G r (z, 1).
In addition to X n,r , we also consider the random variables I n,r and L n,r that model the
number of inner nodes and leaves, respectively, that remain after reducing a random tree
with n nodes r times. The generating functions corresponding to I n,r and L n,r are G r (z, v, 1)
and G r (z, 1, v), respectively.
d d
The relations X n,r = I n,r + L n,r and I n,r = X n,r+1 hold by the combinatorial interpretation of
the operator Φ.

2.3. Asymptotic Analysis. We find explicit generating functions for the factorial moments
of the random variables X n,r , I n,r , and L n,r .
Proposition 2.9. The dth factorial moments of X n,r , I n,r and L n,r are given by
d 1 ∂d 1 ud d!
d
EX n,r = EI n,r−1 = [z n ] G r (z, 1) = [z n ] Ñd−1 (u r )
Cn−1 ∂ vd v=1 Cn−1 (1 + u)(1 − u r+1 )d (1 − u)d−1
(14)
and
1 ud r+2d (1 − u)d! €1Š
d
EL n,r = [z n ] Ñd−1 (15)
Cn−1 (1 + u)(1 − u r+2 )d (1 − u r+1 )d u
where z = u/(1 + u)2 for d ∈ Z≥1 .
Remark. For d ≥ 2, ud Ñd−1 (u−1 ) can be replaced by Ñd−1 (u) in (15), see (5).
Proof. We use the abbreviations
u(1 − u r+1 )2 u r+1 (1 − u)2 1 − u r+2
a := , b := , c := .
(1 − u r+2 )2 (1 − u r+2 )2 (1 − u r+1 )(1 + u)
12 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

We consider the exponential generating function of ∂ d /(∂ v)d G r (z, v) to be a Taylor series
and obtain
X 1 ∂d
G (z, v)q d = G r (z, v + q).
d r
d≥0
d! ∂ v
By Proposition 2.8, extracting the coefficient of q d yields
∂d
G r (z, v) = d![q d ]G r (z, 1 + q) = d!c[q d ]T (a(1 + q), b(1 + q)).
∂ vd v=1

We have
1 − (a − b)(1 + q) − 1 − 2(1 + q)(a + b) + (1 + q)2 (a − b)2
p
T (a(1 + q), b(1 + q)) =
2
1 − (a − b) − (a − b)q
=
2
1 − 2(a + b) + (a − b)2 − 2q(a + b − (a − b)2 ) + q2 (a − b)2
p
− .
2
By using the fact that
(1 − u)(1 − u r+1 )
1 − 2(a + b) + (a − b)2 = ∆2 for ∆=
1 − u r+2
and by choosing α and β such that
a + b − (a − b)2 a−b
α+β = , α−β = ,
∆2 ∆
we obtain

∆€ 1 Æ Š
T (a(1 + q), b(1 + q)) = − (α − β) − (α − β)q − 1 − 2q(α + β) + q2 (α − β)2
2 ∆
∆( ∆1 − 1 − (α − β))
= + ∆T (αq, βq).
2
Extracting the coefficient of q d for d ≥ 1 yields
∂d X €β Š €β Š
G r (z, v) = cd!∆[q d
] α d d
q Ñd−1 = cd!∆α d
Ñd−1
∂ vd v=1
d≥1
α α

where (3) has been used.

Noting that
u u r+1
α= and β=
(1 − u)(1 − u r+1 ) (1 − u)(1 − u r+1 )
completes the proof of (14).
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 13

For the proof of (15), we proceed in the same way and use the identity
1 − ∆ − (a − b)
T (a, b(1 + q)) = + ∆T (α0 q, β 0 q)
2
for
u r+2 u r+1
α =
0
, β =
0
.
(1 − u r+1 )(1 − u r+2 ) (1 − u r+1 )(1 − u r+2 )

From the proof of Proposition 2.9, we extract the following identities for the modified
Narayana polynomials.
Remark. For d ∈ Z≥1 the power series identities
X  n‹ un−d (1 − ux)2n+d−1 (1 − u)d−1 € x(1 − u)2 Š
Ñn−1 = Ñd−1 (x) (16)
n≥1
d (1 − u2 x)2n−1 (1 − ux)2
X un−2d (1 − ux)2n−d−1 (1 − u)2d−1 € x(1 − u)2 Š €1Š
(d)
Ñn−1 = Ñd−1 (17)
n≥1
(1 − u2 x)2n−d−1 d! (1 − ux)2 u
(d)
hold, where Ñn−1 denotes the dth derivative of Ñn−1 .
Proof. In the proof of Proposition 2.9, we showed that
∂d €β Š
cT (av, bv) = cd!∆α d
Ñd−1 .
∂ vd v=1 α
Expanding the left side using (3) and evaluating the derivative yields (16) (where u r has
been replaced by the independent variable x).
The identity (17) is proved in the same way. 
Corollary 2.10. The expected value of X n+1,r is explicitly given by
1 X 2n 2n
 ‹  ‹‹
EX n+1,r = − .
Cn `≥1 n + 1 − `(r + 1) n − `(r + 1)

Proof. Using Proposition 2.9 and Cauchy’s integral formula, we have

u r+1
Cn EX n+1,r = [z n+1 ]
(1 + u)(1 − u r+1 )
I
1 u r+1 dz
=
2πi γ (1 + u)(1 − u ) z n+2
r+1

u (1 − u)(1 + u)2n du
I r+1
1
= ,
2πi γ̃ 1 − u r+1 un+2
where γ is a circle around 0 with a sufficiently small radius such that γ0 , the image of γ under
the transformation, is a small contour circling 0 exactly once as well.
14 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

Expanding (1−u r+1 )−1 into a geometric series and exchanging integration and summation,
we obtain
X
Cn EX n+1,r = [un+1−`(r+1) ](1 − u)(1 + u)2n ,
`≥1

which implies the result. 

Having determined a closed form for this generating function allows us to analyze the
asymptotic behavior of X n,r in a relatively straightforward way.
Theorem 1. Let r ∈ N0 be fixed and consider n → ∞. Then the expected size and the corre-
sponding variance of an r-fold cut plane tree are given by
n r(r − 1)
EX n,r = − + O(n−1 ), (18)
r + 1 6(r + 1)
and
r(r + 2)
VX n,r = n + O(1). (19)
6(r + 1)2
The factorial moments are asymptotically given by
1 d
d
EX n,r = n d
+ (d r 2 − 4d r − 3r 2 − 6d + 6r + 6)nd−1 + O(nd−3/2 )
(r + 1)d 12(r + 1)d
for d ≥ 1. Note that all O-constants above depend implicitly on r.
Proof. In a nutshell, we want to extract the growth of the derivatives of the generating func-
tions ∂∂v d G r (z, 1), as dividing these quantities by Cn−1 yields the factorial moments. We want
d

to extract the growth by means of singularity analysis (cf. [10]).

In order to do so, we first need to establish the location of the dominant singularity of
these generating functions, which are explicitly given in (14).
p The singularities of (14) are roots of unity in terms of u. Substituting back u = (1 −
1 − 4z )/(2z) − 1 maps these roots of unity to real numbers greater or equal to 1/4 and
only u = 1 is mapped to z = 1/4. Thus z = 1/4 is the dominant singularity of (14). A more
detailed treatment of these analytic properties of the substitution z = u/(1+u)2 can be found
in [13, Proposition 2.3].
As N0 (x) = 1, we obtain the expansion
1 1 r2 − r − 3
(1 − u)−1 − + (1 − u) + O((1 − u)2 )
2(r + 1) 4 24(r + 1)
for the function on the right-hand side of (14) with d = 1. Then, the expansion

(1 − u)−κ = 2−κ (1 − 4z)−κ/2 + 2−κ κ(1 − 4z)−(κ−1)/2

κ(κ − 1)
+ 2−κ (1 − 4z)−(κ−2)/2 + O((1 − 4z)−(κ−3)/2 ) (20)
2
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 15

for fixed κ ∈ C yields

1 r2 − r − 3
(1 − 4z)−1/2 + (1 − 4z)1/2 + O((1 − 4z)3/2 ) + power series in (1 − 4z).
4(r + 1) 12(r + 1)
By singularity analysis, the nth coefficient, normalized by Cn−1 , is asymptotically
n r(r − 1)
EX n,r = − + O(n−1 )
r + 1 6(r + 1)
using
1 € 3 −1 Š
Cn−1 = 4n−1p 1 + n + O(n −2
) .
n3/2 π 8
The higher order factorial moments follow similarly by expanding the function on the
right-hand side of (14) for general d > 1 around u = 1 with the help of SageMath, where
in particular the explicit values of the derivatives of the Narayana polynomials from Propo-
sition 2.3 are required.
Singularity analysis of the resulting expansion yields the expression given in the statement
of the theorem. Finally, note that the variance can be computed by using
VX n,r = EX n,r
2
+ EX n,r − (EX n,r )2 .

Theorem 2. The size X n,r of the tree obtained from a random plane tree with n nodes by cutting
it r-times is, after standardization, asymptotically normally distributed for n → ∞ and fixed r,
i.e.,
n
X n,r −
r +1 d
v −→ N (0, 1).
t r(r + 2)
n
6(r + 1)2
To be more precise, for x ∈ R we have
€ X n,r − nµ
Z x
Š 1 2
/2
P p ≤x =p e−t d t + O(n−1/2 ),
σ2 n 2π −∞
1 r(r+2)
with µ = r+1 and σ2 = 6(r+1)2 and where the O-constant depends implicitly on r.
d
As I n,r−1 = X n,r , the same also holds for this random variable.
The rest of this section is devoted to the proof of this central limit theorem. In order to
derive the fact that the number of remaining nodes after r reductions is asymptotically nor-
mally distributed, we first show that the number of nodes that are deleted after r reductions
is asymptotically normally distributed. Then, as the sum of the number of remaining nodes
and the number of deleted nodes is equal to the original tree size, we obtain immediately
that the number of remaining nodes has to be asymptotically normally distributed as well.
16 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

We begin by considering the function F r : T → N0 which maps a plane tree τ to the number
of nodes that are deleted when reducing the tree r times, i.e. the difference between the size
of τ and the size of ρ r (τ). Let τn now denote a plane tree with n nodes.
For the sake of convenience, we consider F r (τn ) to be n if r is larger than the maximal
number of reductions that can be applied to τn before the tree cannot be reduced further. In
particular, this means that F r ( ) = 1 for r ≥ 1.
It is easy to see that the parameter F r (τn ) is a so-called additive tree parameter, meaning
that
F r (τn ) = F r (τi1 ) + · · · + F r (τi` ) + f r (τn )
holds, where τi1 , . . . , τi` are the subtrees rooted at the children of the root of τn , and f r : T →
{0, 1} is a toll function recursively defined by
1, if F r−1 (τik ) = ik for all k = 1, . . . , `,

f r (τn ) =
0, otherwise,
for r ≥ 1 and f0 (τn ) = 0.
In order to prove asymptotic normality for additive tree parameters, we can use [25, Theo-
rem 2], which requires us to show that the expected value of the toll function is exponentially
decreasing in n. This is done in the following lemma.
Lemma 2.11. The expected value of f r (τn ) is exponentially decreasing in n.
Remark. Of course, n − F r (τn ) is also an additive parameter. However, the expected value of
the corresponding growth function is not exponentially decreasing.
Proof. Define
qn,r = E( f r (τn )) = P(F r−1 (τik ) = ik for all k = 1, . . . , `) = P(F r (τn ) = n)
and the corresponding generating function
X
Q r (z) = Cn−1 qn,r z n .
n≥1

Observe that F r (τn ) = n holds if and only if τn has height less than r, as removing all leaves
from a tree reduces its height by precisely one. Therefore, the generating function Q r (z) is
the generating function enumerating trees of height less than r.
It is well-known (cf. [3]) that the generating function for plane trees of height less than r
can be expressed in terms of Fibonacci polynomials as
zF r−1 (−z)
Q r (z) = .
F r (−z)
The roots of F r (−z) are also well-known and can be written as α j,r = (4 cos2 ( jπ/r))−1 for
j = 1, . . . , b(r − 1)/2c.
Thus Q r (z) is a rational function and its coefficients have the form
X
Cn−1 qn,r = c j,r α−n
j,r
j
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 17

for constants c j,r . We have |α j,r | > 4. As

4n−1
Cn−1 ∼ p 3/2 ,
πn
there exists a constant c ∈ (0, 1) such that qn,r = O(c n ). 
Thus, by the strategy discussed above, we find that not only F r (τn ) but also X n,r = n − Fn,r
is asymptotically normally distributed.
Remark. Note that the fact that F1 (τn ) is asymptotically normally distributed means that
the Narayana numbers are asymptotically normally distributed, see for example [7, Theo-
rem 3.13].
As sketched above, Lemma 2.11 allows us to apply [25, Theorem 2] in order to prove that
F r (τn ), and therefore also X n,r = n − F r (τn ) is asymptotically normally distributed. All that
remains to prove is that the speed of convergence is O(n−1/2 ).
We do so by noting that the proof for asymptotic normality in Wagner’s theorem is based
on [7, Theorem 2.23], where a version of Hwang’s Quasi-Power Theorem [16] without
quantification of the speed of convergence is used. Replacing this argument with the multi-
dimensional quantified version given in [15] then gives us the desired speed of convergence
of O(n−1/2 ).

3. CUTTING PATHS
3.1. The Expansion Operator and Results. Let P denote the combinatorial class of paths,
i.e. trees in which every node is either a leaf or has precisely one child. The tree reduction
ρ : T \ P → T which we will focus on in this section reduces a tree by cutting away all paths
of the tree. This operation is illustrated in Figure 4.

7→

FIGURE 4. Illustration of the “cutting paths”-operator ρ

Analogously to our approach in Section 2.2, we first determine the corresponding expan-
sion operator Φ. In order to do so, we need the generating function for the family of paths
P , which is given by P = P(z, t) = 1−z
t
. For the sake of readability, we omit the arguments
of P.
18 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

Proposition 3.1. Let F ⊆ T be a family of plane trees with bivariate generating function
f (z, t), where z marks inner nodes and t marks leaves. Then the generating function for ρ −1 (F ),
the family of trees whose reduction is in F , is given by
z€ zP 2 Š
Φ( f (z, t)) = (1 − P) f , . (21)
(1 − P)2 (1 − P)2
Proof. The fact that Φ is a linear operator is obvious from a combinatorial point of view, mean-
ing that we may concentrate on some tree τ with n inner nodes and k leaves, represented by
zn t k.
We follow the proof of Proposition 2.5 and observe that all possible tree expansions of τ
can be obtained by the following operations: the leaves of τ are expanded by appending
a sequence of at least two paths to each of them. Note that appending a single path to
a leaf is not allowed, because this would just extend the path ending in that leaf, which
causes ambiguity. Then, the inner nodes are expanded as well by appending (possibly empty)
sequences of paths to the 2n + k − 1 available positions between, before, and after their
children.
Translating this expansion to the language of generating functions yields
€ zP 2 Šk 1
Φ(z n t k ) = z n ,
1 − P (1 − P)2n+k−1
which proves (21). 
Corollary 3.2. The generating function for plane trees T (z, t) satisfies the functional equation
T (z, t) = P + Φ(T (z, t)). (22)

Proof. Surjectivity of ρ implies ρ −1 (T ) = T \P , which proves the statement after translating

this into the language of generating functions with the help of Φ. 

In the following proposition, we determine the generating function G r (z, vI , vL ) measuring

the effect of applying the path reduction r times on the size of the tree. Most interestingly,
we will see that the path connection is in fact strongly related to the leaf reduction from the
previous section.

Proposition 3.3. The trivariate generating function G r (z, vI , vL ) = G rP (z, vI , vL ) enumerating

plane trees whose paths can be cut at least r-times, where z marks the tree size and vI and vL
mark the number of inner nodes and leaves of the r-fold cut tree, respectively, is given by
G r (z, vI , vL ) = Φ r (T (zvI , t vL ))| t=z
r+1 r+1 r+1
u(1 − u2 −1 )2 u2 −1 (1 − u)2

1 − u2
‹
= T vI , vL ,
(1 − u2r+1 −1 )(1 + u) (1 − u2r+1 )2 (1 − u2r+1 )2

where z = u/(1 + u)2 .

 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 19

Proof. By the same reasoning as in the proof of Proposition 2.8, the generating function we
are interested in is G r (z, vI , vL ) = Φ r (T (zvI , t vL ))| t=z , meaning that we want to study the
iterated application of Φ. To do so, we consider the strongly related operator
€ z zP 2 Š
Ψ( f (z, t)) := f , .
(1 − P)2 (1 − P)2
The relation
r−1
Y
Φ ( f (z, t)) = Ψ ( f (z, t))
r r
(1 − Ψ j (P))
j=0

can be proved easily by induction and enables us to determine the behavior of Φ via Ψ.
First of all, for r ≥ 0 and r ≥ 1, the relations
z z(Ψ r−1 (P))2
Ψ r (z) = Q r−1 and Ψ r (P) = Q r−1
j=0
(1 − Ψ j (P))2 j=0
(1 − Ψ j (P))2 − z
can be proved easily by induction, respectively. Also observe that we can write Ψ r (t) =
Ψ r (z)Ψ r−1 (P)2 . Now let f r = Ψ r (z)| t=z , g r = Ψ r (t)| t=z , and h r = Ψ r (P)| t=z . With the help of
Qr j 2 r+1
the identity j=0 (1 + u2 ) = 1−u 1−u we are able to prove the explicit formula
r+1 r+1
u2 −1 (1 − u) z 2 −1
hr = = , (23)
1 − u2r+2 −1 F2r+2 −1 (−z)
where z = u/(1 + u)2 and the second equation is a consequence of (8). Using (23), we
immediately find
r+1 r+1 r+1
u(1 − u2 −1 )2 zF2r+1 −1 (−z)2 u2 −1 (1 − u)2 z 2 −1
fr = = and gr = = .
(1 − u2r+1 )2 F2r+1 (−z)2 (1 − u2r+1 )2 F2r+1 (−z)2
Putting everything together yields
r+1
z n+(2 F2r+1 −1 (−z)2n−1
−1)k
Φ (z t )| t=z =
r n k
F2r+1 (−z)2n+2k−1
r+1
1 − u2 € u(1 − u2r+1 −1 )2 Šn € u2r+1 −1 (1 − u)2 Šk
= ,
(1 − u2r+1 −1 )(1 + u) (1 − u2r+1 )2 (1 − u2r+1 )2
which directly implies the statement. 
The following result shows that there is an intimate connection between the “cutting
leaves”-reduction from Section 2 and the “cutting paths”-reduction, as can be seen after com-
paring the statement of Proposition 2.8 with the statement of Proposition 3.3.
Corollary 3.4. The generating function G r (z, vI , vL ) = G rP (z, vI , vL ) measuring the change in
size after cutting away all paths from plane trees r times is equal to the generating function
G2Lr+1 −2 (z, vI , vL ) measuring the change in size after cutting away all leaves from plane trees
2 r+1 − 2 times.
20 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

This connection is now especially important for the analysis of the random variable X n,r =
P
X n,rmodeling the number of nodes that are left after reducing a random tree τ with n nodes
r times by removing all paths. In fact, it follows that
d
P
X n,r = X n,2
L
r+1 −2 ,

P
meaning that the asymptotic analysis of the factorial moments of X n,r as well as the limiting
distribution follow directly from the corresponding results in Section 2.3.
Theorem 3. Let r ∈ N0 be fixed and consider n → ∞. Then expectation and variance of the
random variable X n,r = X n,r
P
can be expressed as
n (2 r − 1)(2 r+1 − 3)
EX n,r = − + O(n−1 ), (24)
2 −1
r+1 3(2 − 1)
r+1

and
2 r+1 (2 r − 1)
VX n,r = n + O(1). (25)
3(2 r+1 − 1)2
The factorial moments are asymptotically given by
nd
d
EX n,r =
(2 r+1 − 1)d
d
+ (4 r+1 d − 2 r+4 d − 3 · 4 r+1 + 9 · 2 r+2 + 6d − 18)
12(2 − 1)d
r+1

+ O(nd−3/2 ).
Furthermore, X n,r = X n,r
P
is asymptotically normally distributed, i.e., for x ∈ R we have
Z x
X n,r − µn
 
1 2
P p ≤x =p e−t /2 d t + O(n−1/2 )
σ2 n 2π −∞
1 2 r+1 (2 r −1)
for µ = 2 r+1 −1 and σ2 = 3(2 r+1 −1)2 . All O-constants in this theorem depend implicitly on r.
3.2. Total number of paths. In the context of this reduction it is interesting to investigate
the total number of paths needed to construct a given tree. To determine this parameter
we can reduce the tree repeatedly and count the number of leaves. The sum of the number
of leaves over all reduction steps is equal to the number of paths, which follows from the
observation that leaves mark the endpoints of all paths.
Formally, given the random variables Pn,r counting the number of leaves
P in the rth reduc-
tion of a tree of size n, we want to analyze the random variable Pn := r≥0 Pn,r .
Proposition 3.5. The expected number of paths needed to construct a uniformly random tree
of size n satisfies
r
1 n 1−u
X u2
EPn = [z ] , (26)
Cn−1 1 + u r≥1 (1 − u2r )(1 − u2r −1 )
where z = u/(1 + u)2 .
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 21

Proof. As a consequence of Proposition 3.3, the bivariate generating function enumerating

plane trees where z marks tree size and v marks the number of leaves after r path reductions
can be written as
r+1
1 − u2 € u(1 − u2r+1 −1 )2 u2r+1 −1 (1 − u)2 Š
T , v .
(1 − u2r+1 −1 )(1 + u) (1 − u2r+1 )2 (1 − u2r+1 )2
By differentiating this generating function once with respect to v and setting v = 1 after-
wards, we obtain an expression where Cn−1 EPn,r can be extracted as the coefficient of z n .
By (15) with d = 1 and r replaced by 2 r+1 − 2, we have
r+1
1 1−u u2
EPn,r = [z ]
n
.
Cn−1 1 + u (1 − u2r+1 )(1 − u2r+1 −1 )
Summation over r ≥ 0 and shifting the index of summation by one completes the proof. 
Our strategy for determining an asymptotic expansion for EPn as given in (26) is based on
the Mellin transform.
Theorem 4. For n → ∞, the expected number of paths required to construct a uniformly
random tree of size n is given by the asymptotic expansion
1 γ + 4(α − 1) log 2 + log 2 + 24ζ0 (−1) + 2
EPn = (α − 1)n + log4 n − + δ(log4 n) + O(n−1/4 ),
6 12 log 2
(27)
where
1 X
δ(x) := (−1 + χk )Γ (χk /2)ζ(−1 + χk )e2kπi x (28)
log 2 k∈Z\{0}
with χk = 2kπi α = k≥1 1/(2 − 1) ≈ 1.606695, γ is the
P k
log 2 is a fluctuation with mean 0 and :
Euler–Mascheroni constant and ζ is the Riemann zeta function.
Remark. The constant α appears in the asymptotic analysis of digital search trees (see e.g. [20]).
Proof. In order to obtain an asymptotic expansion from (26), we rewrite
r r r
1−u X u2 u X € u2 −1 u2 Š
P(z) = = −
1 + u r≥1 (1 − u2r )(1 − u2r −1 ) 1 + u r≥1 1 − u2r −1 1 − u2r
where z = u/(1 + u)2 . The main task to obtain an asymptotic expansion of P(z) is to provide
a precise analysis of this sum, which we carry out via the Mellin transform. We consider the
function
X e−(2r −1)t X e−2r t
f (t) := − ,
r≥1
1 − e−(2r −1)t r≥1 1 − e−2r t
obtained from substituting u = e−t in the sum above. With
X 1 X ` + s − 1‹ 1
A(s) := ((1 − 2 ) − 1) =
−r −s

r≥1
2 rs
`≥1
` 2s+` − 1
22 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

we find that the corresponding Mellin transform of this difference of harmonic sums is given
by
f ∗ (s) = Γ (s)ζ(s)A(s)
with fundamental strip 〈1, ∞〉. In order for the inversion formula to be valid, we need to
show that f ∗ (s) decays sufficiently fast along vertical lines in the complex plane. While Γ (s)
and ζ(s) are well-known to decay exponentially and grow polynomially along vertical lines,
respectively, the Dirichlet series A(s) has to be investigated in more detail.
We want to estimate the summands in
s X 1 € s Š
A(s) − s+1 = (1 − 2 )
−r −s
− 1 − .
2 − 1 r≥1 2 rs 2r

To do so, we consider g(x) = (1 − x)−s as a function of a real variable. By means of the

integral form of the Taylor approximation error we find
Z 2−r
|g(2−r ) − g(0) − g 0 (0) · 2−r | = s(s + 1)(1 − t)−s−2 (2−r − t) d t
0
Z 2−r

≤ |s||s + 1|2−r |1 − t|− Re s−2 d t

0
≤ |s||s + 1|2 −2r
(1 − 2−r )− Re s−2 ,
where the last inequality is valid under the assumption that Re s > −2. Using this estimate,
we find
s s
|A(s)| ≤ A(s) − s+1 + s+1
2 −1 2 −1
s X 1
≤ s+1 + |s||s + 1| ,
2 −1 r≥1
(2 − 1)Re s+2
r

where the sum converges for Re
s > −2. Therefore, A(s) has polynomial growth in Im s for
2πi 1
Re s > −2 and Im s = log 2 k + 2 , where k ∈ Z and |k| → ∞, as well as on vertical lines
with Re s > −2 and Re s 6= −1. This implies that f ∗ (s) decays sufficiently fast, and thus the
inversion formula states
Z 2+i∞
1
f (t) = Γ (s)ζ(s)A(s)t −s ds, (29)
2πi 2−i∞
which is valid for real, positive t → 0 (and thus u → 1− and z → (1/4)− , as we have z =
u/(1+u)2 and u = e−t ). In order to extract the coefficient growth (in terms of z) with the help
of singularity analysis, we require analyticity in a larger region (cf. [10]), e.g. in a complex
punctured neighborhood of 1/4 with1 |arg(z − 1/4)| > 2π/5.
1
Note that the bound 2π/5 is somewhat arbitrary: the argument just needs to be less than π/2.
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 23

Substituting back t for z, we find

€ 1 − p1 − 4z Š p 2
t = − log − 1 = 2 1 − 4z + (1 − 4z)3/2 + O((1 − 4z)5/2 ),
2z 3
which implies
1
|arg t| = |arg(1 − 4z)| + o(1)
2
such that we have the bound |arg t| < 2π/5 for t → 0, given that the restriction on the
argument in terms of z is satisfied.
With the help of our estimates on f ∗ (s) that we discussed above, we find that
€ € 𠊊
| f ∗ (s)t −s | = O |Im(t)|4 |t|− Re(s) exp − |Im(s)| (30)
10
2πi 1


for −3/2 ≤ Re s ≤ 2 and Im s = log 2 k + 2 , where k ∈ Z and |k| → ∞. This is a consequence
of combining the quantified growth of Γ (s) (see [6, 5.11.3]) and the growth of ζ(s) (see [26,
13.51]) with the facts that A(s) is of order O(Im(s)2 ) and 2s+1s −1 is of order O(Im(s)) for s
taking values in the specified region.
We can evaluate (29) by shifting the line of integration from Re(s) = 2 to Re(s) = −3/2
and collecting the residues of the poles we cross. This yields
Z −3/2+i∞
X 1
f (t) = Ress=p ( f (s)t ) +
∗ −s
f ∗ (s)t −s ds,
p∈P
2πi −3/2−i∞
where P = {−1, 1} ∪ {−1 + χk | k ∈ Z \ {0}}. For the error term we use the estimate above
and find Z −3/2+i∞
1
f ∗ (s)t −s ds = O(|t|3/2 ).
2πi −3/2−i∞
Evaluating the residues yields
1 log 2 + 2γ + 24ζ0 (−1)
f (t) = A(1)t −1 + t log t + t
12 log 2 24 log 2
X 1
+ Γ (χk )ζ(−1 + χk )t 1−χk + O(|t|3/2 ).
k∈Z\{0}
log 2

Note that with α := k≥1 1/(2k − 1), we have A(1) = α − 1.

P

When substituting back in order to obtain an expansion in terms of z → 1/4, we have to

carefully check that the error terms within the sum of the residues at χk for k ∈ Z \ {0} can
still be controlled. Considering that for some exponent κ, we have the expansion
t −κ = (1 − 4z)−κ/2 (1 + O(1 − 4z))−κ/2 ,
and thus
€ κ Š
|(1 + O(1 − 4z))−κ/2 − 1| = exp − log(1 + O(1 − 4z)) − 1
2
24 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

κ € κ Š
≤ |log(1 + O(1 − 4z))| exp |log(1 + O(1 − 4z))|
2 2
κ € κ Š
= O(1 − 4z) exp O(1 − 4z) .
2 2
Setting κ = −1+χk shows that the errors that we sum are of order O(|k|(1−4z) exp(|k|O(1−
4z))). Choosing z sufficiently close to 1/4 ensures that the exponential growth is negligible
compared to the exponential decay proved in (30). p
u
Finally, it is easy to see that the factor 1+u can be rewritten as 1− 21−4z . Multiplying our
expansion of f (t) with this factor and substituting back yields the expansion

α−1 α−1 1
P(z) = (1 − 4z)−1/2 − − (1 − 4z)1/2 log(1 − 4z)
4 4 24 log 2
2γ − 2α log 2 + 5 log 2 + 24ζ0 (1)
+ (1 − 4z)1/2
24 log 2
1 X
+ Γ (χk )ζ(−1 + χk )(1 − 4z)(1−χk )/2 + O((1 − 4z)3/4 ).
log 2 k∈Z\{0}

Applying singularity analysis, normalizing the result by Cn−1 , and rewriting the coefficients
of the contributions from the poles at −1 + χk via the duplication formula for the Gamma
function (cf. [6, 5.5.5]) then proves the asymptotic expansion for EPn . 

4. CUTTING OLD LEAVES

4.1. Preliminaries. In this section we consider a slightly more complex reduction: instead
of removing all leaves, we just remove all leftmost leaves. Following [2], we call a leaf that
is a leftmost child an old leaf.
In order to describe the corresponding expansion in the language of generating functions,
we need to change our underlying combinatorial model of trees in a way that specifically
marks old leaves.
Let L be the combinatorial class of plane trees where „ marks old leaves and marks all
nodes that are neither old leaves nor parents thereof. Now, as a first step we determine the
bivariate generating function L(z, w) of L .

Proposition 4.1. The generating function L(z, w) enumerating plane trees with respect to old
leaves „ (marked by the variable w) and all nodes that are neither old leaves nor parents
thereof (marked by z) is given by
p
1 − 1 − 4z − 4w + 4z 2
L(z, w) = . (31)
2
n−2


For n ≥ 2 there are Ck−1 n−2k 2n−2k plane trees of size n (meaning n nodes overall) with k old
leaves.
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 25

For example in Figure 5, the original tree corresponds to z 3 w3 because it has three old
leaves (dashed nodes) and three nodes which are neither old leaves nor parents of old leaves.

7→ 7→ 7→ · · · 7→

FIGURE 5. Illustration of the “cutting old leaves”-operator ρ

Proof. We consider the symbolic equation describing the combinatorial class L of plane trees
with respect to old leaves, which is illustrated in Figure 6. The functional equation that can

= + +
P P
L
n≥0 n≥0

L L ··· L L− L L ··· L

n n

FIGURE 6. Symbolic equation for plane trees w.r.t. old leaves

be derived from the symbolic equation by marking „ with w and with z is

w + z(L(z, w) − z)
L(z, w) = z + . (32)
1 − L(z, w)
Solving this equation and choosing the correct branch of the root yields (31).
To extract coefficients of L(z, w), we rewrite it as
v
1
 t 4w
‹
1
 X 1/2‹ (−1)k 4k wk ‹
L(z, w) = 1 − (1 − 2z) 1 − = 1− (33)
2 (1 − 2z)2 2 k≥0
k (1 − 2z)2k−1
wk n + 2k − 2 n k n
X X  ‹
=z+ Ck−1 =z+ Ck−1 2 w z . (34)
k≥1
(1 − 2z)2k−1
k≥1
n
n≥0


26 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

As we will see in the next section, the polynomials defined below will play a similar role for
the “old leaves”-reduction as the Fibonacci polynomials played for the “leaves”- and “paths”-
reduction.
Definition. The polynomials B r (z) are the generating functions of binary trees w.r.t. the num-
ber of internal nodes of height ≤ r satisfying
B r (z) = 1 + zB r−1 (z)2 (35)
for r ≥ 1 and B0 (z) = 1.
4.2. The Expansion Operator and Asymptotic Results. As described in the previous sec-
tion, we now concentrate on the reduction ρ : L → L , which removes all old leaves from a
tree. Note that ρ( ) = , as the root itself is not an old leaf. We begin our analysis of this
reduction by determining the expansion operator Φ.
Proposition 4.2. Let F ⊆ L be a family of plane trees with bivariate generating function
f (z, w), where z marks nodes that are neither old leaves nor parents thereof and w marks old
leaves. Then the generating function for ρ −1 (F ), the family of trees whose reduction is in F , is
given by
Φ( f (z, w)) = f (z + w, (2z + w)w). (36)
Proof. Linearity of Φ is obvious from the combinatorial interpretation, meaning that we can
focus on the expansion of any tree represented by z n wk , i.e. a tree with n nodes that are
neither old leaves nor parents thereof and k old leaves.
Figure 7 illustrates all three possibilities to expand an old leaf „:
– appending an old leaf to the parent of „, which turns the original old leaf into ,
– appending an old leaf to „ itself, which turns the parent into ,
– appending old leaves both to „ and its parent.

Φ
−
→ + +

FIGURE 7. All possible expansions of an old leaf

In terms of generating functions, this means that w is substituted by 2zw + w2 .

Furthermore, the nodes represented by can optionally be expanded by attaching an old
leaf to them, otherwise they stay as they are. This option corresponds to the substitution
z 7→ z + w.
There are no more operations to expand the tree, so putting everything together yields
Φ(z n wk ) = (z + w)n (2zw + w2 )k ,
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 27

which proves the statement. 

An immediate consequence of the fact that ρ : L → L is surjective is the following corol-
lary.
Corollary 4.3. The generating function for plane trees L(z, w) satisfies the functional equation
Φ(L(z, w)) = L(z, w).
We now focus on determining the generating function measuring the change in the tree
size after repeatedly applying the reduction ρ.
Proposition 4.4. Let r ∈ N0 . The bivariate generating function G r (z, v) = G rOL (z, v) enumerat-
ing plane trees, where z marks the tree size and v marks the size of the r-fold cut tree, is given
by
G r (z, v) = Φ r (L(zv, wv 2 ))|w=z 2 = L(zB r (z)v, z(B r+1 (z) − B r (z))v 2 ),
where the B r (z) are the polynomials enumerating binary trees of height ≤ r w.r.t. the number
of internal nodes.
Proof. First, note that the size of a tree with k old leaves and n nodes that are neither old
leaves nor parents thereof is actually n+2k, as parents of old leaves are not explicitly marked.
This explains why we have to substitute w = z 2 in order to arrive at the tree size.
In contrast to the previous sections, the operator Φ is already linear and multiplicative,
meaning that we have
Φ r (z n wk ) = Φ r (z)n Φ r (w)k .
Investigating the repeated application of Φ to z and w leads to the recurrences
Φ r (z) = Φ r−1 (z) + Φ r−1 (z)2 − Φ r−2 (z)2 and Φ r (w) = Φ r+1 (z) − Φ r (z)
for r ≥ 2 and r ≥ 0, respectively. With the recurrence for the polynomials B r from (35) it is
easy to prove by induction that
Φ r (z)|w=z 2 = zB r (z)
for r ≥ 0. Thus, we also find Φ r (w)|w=z 2 = z(B r+1 (z) − B r (z)). Overall, we obtain
Φ r (z n wk )|w=z 2 = z n+k B r (z)n (B r+1 (z) − B r (z))k ,
which, by linearity of Φ, proves the proposition. 
For the next step in our analysis, we turn to the random variable X n,r = X n,r
OL
which models
the size of the tree that results from reducing a random tree τ with n nodes r-times.
As we have ρ( ) = (and thus no trees vanish completely), the probability generating
function for this random variable is simply
[z n ]G r (z, v)
Ev X n,r
= .
Cn−1
While the height polynomials B r (z) make it very difficult to obtain general results for the
factorial moments of X n,r , special moments like expectation and variance are no problem,
and even a central limit theorem is possible.
28 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

Theorem 5. Let r ∈ N0 be fixed and consider n → ∞. Then the expected tree size after deleting
old leaves of a tree with n nodes r-times and the corresponding variance are given by
B r0 (1/4)
EX n,r = (2 − B r (1/4))n − + O(n−1 ), (37)
8
and
(2 − B r (1/4))B r0 (1/4)
 
VX n,r = B r (1/4) − B r (1/4)2 + n + O(1). (38)
2
All O-constants in this theorem depend implicitly on r.
Additionally, the random variable X n,r is asymptotically normally distributed for fixed r ≥ 1,
i.e.
X n,r − µn d
p −→ N (0, 1),
σ2 n
(2−B r (1/4))B r0 (1/4)

where µ = (2 − B r (1/4)) and σ2 = B r (1/4) − B r (1/4)2 + 2 .
Proof. First of all, we observe that Proposition 4.1 and Proposition 4.4 combined with the
recursion B r (z) = 1 + zB r−1 (z)2 allow us to write the bivariate generating function as

1 − 1 − 4zv(B r (z)(1 − v) + v)
p
G r (z, v) = .
2
The asymptotic expansion for the expected value EX n,r can now be obtained by determining
1 ∂ 1 z(2 − B r (z))
[z n ] G r (z, v)| v=1 = [z n ] p .
Cn−1 ∂v Cn−1 1 − 4z
By means of singularity analysis we find
B r0 (1/4) € 3B 0 (1/4)
r
3B r00 (1/4) Š
EX n,r = (2 − B r (1/4))n − − + n−1 + O(n−2 ),
8 16 128
which proves (37). For the second factorial moment we obtain
1 ∂2 1 € 2z 2 (2 − B (z)) 2z(1 − B (z)) Š
r r
2
EX n,r = [z ] 2 G r (z, v)| v=1 =
n
[z ]
n
+ ,
Cn−1 ∂v Cn−1 (1 − 4z) 3/2 (1 − 4z)1/2
which yields
€ (2 − B r (1/4))B r0 (1/4) Š
2
EX n,r = (2 − B r (1/4))2 n2 + 2B r (1/4) − B r (1/4)2 − 2 + n
4
(2 − B r (1/4))B r00 (1/4) B r0 (1/4)2 B r (1/4)B r0 (1/4)
+ − − + O(n−1 ).
64 64 8
2
The variance can now be obtained via VX n,r = EX n,r + EX n,r − (EX n,r )2 , which proves (38).
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 29

In order to show asymptotic normality of X n,r we investigate the random variable n − X n,r ,
which counts the number of nodes that are deleted after reducing some tree r times. Observe
that this quantity can be seen as an additive tree parameter F r defined recursively by
F r (τn ) = F r (τi1 ) + F r (τi2 ) + · · · + F r (τi` ) + f r (τn ) and Fr ( ) = 0
where τn is some tree of size n, τi1 up to τi` are the subtrees rooted at the children of the
root of τn , and f r : L → {0, 1, . . . , r − 1} is a toll function defined by
r−1 
X 1 if ρ j (τn ) has an old leaf attached to its root,
f r (τn ) =
j=0
0 otherwise,

for r ≥ 1. Now, as f r (τn ) enumerates the number of old leaves deleted from the root of τn
after r reductions, F r (τn ) equals the total number of deleted nodes after r reductions.
The fact that r is fixed implies that f r is not only bounded, but also a so-called local func-
tional, meaning that the value of f r (τn ) can already be determined from the first r levels of
τn . This is because one application of ρ can reduce the distance between the root of the tree
and the closest old leaf by at most one. Thus all old leaves that are deleted from the root
during r reductions have to be found within the first r levels of τn .
As we have now established that f r is both bounded and a local functional, we are able to
apply [18, Theorem 1.13], which proves that n − X n,r is asymptotically normally distributed.
Thus X n,r is asymptotically normally distributed as well, which proves the statement. 
Remark. In [9], the asymptotic behavior of a sequence strongly related to B r (1/4) was stud-
B (1/4)
ied: in Section 4, the authors define a sequence f n such that f r+1 = 21 − r 4 , in our notation.
They prove the asymptotic expansion f n = n+log 1n+O(1) . This allows us to conclude that the
asymptotic behavior of B r (1/4) can be described as
4 4 log r
B r (1/4) = 2 − + 2
+ O(r −2 )
r r
for r → ∞.

5. CUTTING OLD PATHS

5.1. The Expansion Operator. As in previous sections, we adapt the “old leaves” reduction
to remove all “old paths”. That is, the tree reduction ρ : L → L in this section reduces a tree
by removing all paths that end in an old leaf. This operation is illustrated in Figure 8, where
„ marks old leaves and marks all nodes that are neither old leaves nor parents thereof.
Obviously, we also need the combinatorial class of paths P for our analysis. The bivariate
generating function of P is given by P = P(z, w) = 1−z w
, where w and z mark „ and ,
respectively. Also, we omit the arguments of P for the sake of readability. Now, we determine
the shape of the expansion operator Φ.
Proposition 5.1. Let F ⊆ L be a family of plane trees with bivariate generating function
f (z, w), where z marks nodes that are neither old leaves nor parents thereof and w marks old
30 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

7→ 7→

FIGURE 8. Illustration of the “cutting old paths”-operator ρ

leaves. Then the generating function for ρ −1 (F ), the family of trees whose reduction is in F , is
given by
Φ( f (z, w)) = f (z + P, zP + P 2 ). (39)
Proof. With linearity of the operator Φ being obvious from a combinatorial point of view, we
only have to investigate the expansion of any tree represented by z n wk , i.e. a tree with n
nodes that are neither old leaves nor parents thereof and k old leaves.
There are two options to expand an old leaf „:
– either appending an old path to the parent of „, which turns the old leaf into ,
– or an old path is appended to both the parent of „ and to „ itself.
Note that just appending an old path to „ is not a valid expansion as this introduces ambiguity.
This is the same argument that we also used in the proof of Proposition 3.1. Overall, this
means that Φ has to map w to zP + P 2 .
On the other hand, the nodes represented by can optionally be expanded by attaching
an old path. Otherwise they stay as they are. Overall, this implies Φ(z) = z + P.
Putting everything together, we immediately arrive at the statement of the Proposition. 
Analogously to the previous reductions, surjectivity of ρ : L → L implies the following
corollary.
Corollary 5.2. The generating function for plane trees L(z, w) satisfies the functional equation
Φ(L(z, w)) = L(z, w).
In order to carry out a detailed analysis of this reduction, we need information about the
iterated application of Φ to L(zvI , wvL2 ), which leads to the generating function G r (z, vI , vL2 )
measuring the change in the tree size after r applications of the reduction. The following
proposition deals with determining this generating function.
Proposition 5.3. Let r ∈ N0 . The trivariate generating function G r (z, vI , vL2 ) = G rOP (z, vI , vL2 )
enumerating plane trees, where z marks the tree size, vL marks all old leaves, and vI marks all
nodes that are neither old leaves nor parents thereof, is given by
u(1 − u r+1 ) u r+2 (1 − u)2
 ‹
G r (z, vI , vL ) = Φ (L(zvI , wvL ))|w=z 2 = L
2 r 2
vI , 2
v ,
(1 + u)(1 − u r+2 ) (1 + u)2 (1 − u r+2 )2 L
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 31

where z = u/(1 + u)2 .

Proof. Observe that the operator Φ is already linear and multiplicative, which is why we can
concentrate on finding suitable expressions for Φ r (z) and Φ r (w).
First of all, for r ≥ 1 the recurrences
Φ r (z) = Φ r−1 (z) + Φ r−1 (P), Φ r (w) = Φ r−1 (P)Φ r (z)
follow immediately from (39). Furthermore, the relation
r
Y Φ j (z)
Φ (P) = P
r

j=1
1 − Φ j (z)

can easily be proved by induction. Then, by setting f r := Φ r (z)|w=z 2 the recurrences above
translate to
r−1
Y fj
f r = f r−1 + z .
j=0
1 − fj
As a next step, we show by induction that f r can be expressed in terms of Fibonacci polyno-
mials as
zF r+1 (−z)
fr = ,
F r+2 (−z)
where in particular (7) was used. As a consequence, we find
zF r+2 (−z) zF r+1 (−z) z r+2
Φ r (P)|w=z 2 = f r+1 − f r = − = .
F r+3 (−z) F r+2 (−z) F r+2 (−z)F r+3 (−z)
This allows us to express g r := Φ r (w)|w=z 2 as
z r+2
g r = Φ r−1 (P)|w=z 2 · f r = .
F r+2 (−z)2
Finally, as we have Φ r (z n wk )|w=z 2 = f rn g rk , substituting z = u/(1+u)2 and using (8) completes
the proof. 

5.2. Analysis of Tree Size and Related Parameters. We investigate the behavior of the
random variable X n,r = X n,r
OP
which models the number of nodes remaining after reducing
a random tree τ with n nodes r-times. The tree τ is chosen uniformly among all trees
of size n. Analogously to the “old leaf”-reduction from the previous section, we also have
ρ( ) = for the “old path”-reduction, meaning that no trees vanish completely. For the sake
of convenience we set G r (z, v) := G r (z, v, v 2 ), allowing us to write the probability generating
function of X n.r as
[z n ]G r (z, v)
Ev X n,r = .
Cn−1
32 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

With the help of Proposition 5.3, it is easy to obtain expressions for the factorial moments
d
EX n,r for fixed d by differentiating G r (z, v) d-times with respect to v and setting v = 1 af-
terwards. General expressions for d ≥ 2 (coinciding with the value given for d = 2) are
available but less pleasant.
Lemma 5.4. The factorial moments of X n,r are
1 u(1 + u r+1 )
EX n,r = [z ]
n
,
Cn−1 (1 + u)(1 − u r+2 )
2 (1 + u)u r+2
EX n,r (X n,r − 1) = [z n ]
Cn−1 (1 − u)(1 − u r+2 )2
and
v
d! n 1 − u u(1 + u r+1 ) u t 1 − ur d
 ‹
d
EX n,r = [z ] +
Cn−1 1 + u (1 − u)(1 − u r+2 ) 1 − u 1 − u r+2
v
− u r+2 + 2u r+1 − u r + 2 2(1 + u r+1 )(1 − u r+2 ) t 1 − u r
 2r+2
2u
‹
× Ñd−1 +
(1 + u)2 u r u r (1 + u)2 1 − u r+2
for d ≥ 2.
Proof. The expressions for d ∈ {1, 2} can be obtained by differentiation. We consider the
general case here.
We use the abbreviations
u(1 − u r+1 ) u r+2 (1 − u)2 1−u
a= , b = , ∆ = .
(1 + u)(1 − u r+2 ) (1 + u)2 (1 − u r+2 )2 1+u
By the same argument as in the proof of Proposition 2.9, we have
∂d
G r (z, v) = d![q d ]L(a(1 + q), b(1 + q)2 ).
∂ vd v=1

By using (31), we rewrite L(a(1 + q), b(1 + q)2 ) as

1 − (1 − 2a − 2aq)2 − 4b(1 + q)2
p
L(a(1 + q), b(1 + q) ) =
2
2
1 − (1 − 2a) − 4b − 2q(2a(1 − 2a) + 4b) + q2 (4a2 − 4b)
p
2
= .
2
We have v v
t 4a2 − 4b 2u t 1 − ur
(1 − 2a)2 − 4b = ∆2 , = .
∆2 1 − u 1 − u r+2
We choose α and β such that
v
2a(1 − 2a) + 4b t 4a2 − 4b
α+β = , α − β = .
∆2 ∆2
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 33

This results in
1
− 1 + q(α − β) 1 − q(α − β) − 1 − 2q(α + β) + q2 (α − β)
p
∆
L(a(1 + q), b(1 + q) ) = ∆
2
+∆
2 2
1
∆ − 1 + q(α − β)
=∆ + ∆T (αq, βq).
2
Using (3) to extract the coefficient of q d for d ≥ 1 yields
∂d €α − β € ⠊Š
G r (z, v) = d!∆ [d = 1] + α d
Nd−1 .
∂ vd v=1 2 α
Inserting everything concludes the proof of the proposition. 
Corollary 5.5. The expected value of X n+1,r is explicitly given by
1 2n 2n 2n
 ‹ X ‹  ‹‹‹
EX n+1,r = + −
Cn n j≥0
n − ( j + 1)(r + 2) + 1 n − j(r + 2) − 1

Proof. From Lemma 5.4, we obtain

(1 + u r+1 )u
Cn EX n+1,r = [z n+1 ]
(1 + u)(1 − u r+2 )
and proceeding as in Corollary 2.10 we obtain the given result. 
By expanding the expressions in Lemma 5.4 and using singularity analysis, we obtain the
asymptotic growth of the expected value and the variance.
Theorem 6. Let r ∈ N be fixed and consider n → ∞. Then the expected size and the corre-
sponding variance of an r-fold cut plane tree are given by
2n r(r + 1)
EX n,r = − + O(n−1 ),
r + 2 3(r + 2)
and
2r(r + 1)
VX n,r = n + O(1).
3(r + 2)2
For d ≥ 3, the dth factorial moment is
p
2d−1 d 2d − 5
 ‹
rπ d d− 21
d
EX n,r= n d
+ n + O(nd−1 ).
(2d − 3)(r + 2)d d − 2 2d−3 (r + 2)d−1/2
All O-constants in this theorem depend implicitly on r.
Besides the analysis of the tree size, we are also interested in how the numbers of nodes
represented by „ and by develop when the tree is reduced repeatedly. Formally, this means
that we consider the random variables X n,r„
and X n,r counting the number of old leaves and
the number of all nodes that are neither old leaves nor parents thereof, respectively. By
construction, the relation
X n,r = 2 · X n,r
„
+ X n,r (40)
34 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

holds.
The bivariate generating functions corresponding to these random variables can be ob-
tained directly from Proposition 5.3. We have
G r„ (z, v) = G r (z, 1, v), G r (z, v) = G r (z, v, 1).
In contrast to X n,r , the dth factorial moments for X n,r
„
and X n,r have simpler expressions.

Proposition 5.6. Let d ∈ N. Then the dth factorial moments of X n,r

„
and X n,r are given by

d (2d − 2)d−1 n 1 − u u r d+2d

„
EX n,r = [z ] (41)
Cn−1 1 + u (1 − u r+2 )2d
and
1 u(1 − u r+1 )(1 + u r+2 )
EX n,r = [z n ] (42)
Cn−1 (1 + u)(1 − u r+2 )2
as well as
d 1 (1 − u r+1 )d ud 2d d!
EX n,r = [z ] n
Ñd−1 (u r+2 ) (43)
Cn−1 (1 − u) (1 + u)(1 − u )
d−1 r+2 2d

for d > 1.
Proof of Proposition 5.6. As in the proof of Lemma 5.4, we use the abbreviations
u(1 − u r+1 ) u r+2 (1 − u)2 1−u
a= , b= , ∆= .
(1 + u)(1 − u r+2 ) (1 + u)2 (1 − u r+2 )2 1+u
Then, using (33), we get
∂d „ ∂d 1 − 2a € 1 Šd € 4bv Š1/2−d (−4b)d
G (z, v) = L(a, bv) = − 1 − .
∂ vd r ∂ vd 2 2 (1 − 2a)2 (1 − 2a)2d
Setting v = 1 and using the fact that
(1 − 2a)2 − 4b = ∆2 (44)
proves (41).
For deriving ∂ d /(∂ v)d G r (z, v), we proceed as in the proof of Proposition 2.9. The crucial
identity is
1
− 1 + (α0 − β 0 )q
L(a(1 + q), b) = ∆ ∆ + ∆T (α0 q, β 0 q)
2
with
2u(1 − u r+1 ) β0
α0 = , = u r+2 .
(1 − u r+2 )2 (1 − u) α0
This implies (42) and (43).

As in Section 2.3, the above proof exhibits some identities:
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 35

Remark. For d ∈ Z≥1 , the power series identities

X un+k x k (1 − x)n (1 − u)2k 
n + 2k − 2 n
‹
ud x d (1 − u)
k d
C k−1 2 = (2d − 2) d−1
(45)
n≥0
(1 + u)n+2k (1 − ux)n+2k n (1 − ux)2d (1 + u)
k≥1

and
X un+k x k (1 − x)n (1 − u)2k n + 2k − 2 n (1 − x)d ud 2d d! Ñd−1 (ux)
 ‹
d
n Ck−1 2 = (46)
n≥0
(1 + u)n+2k (1 − ux)n+2k n (1 − u)d−1 (1 + u)(1 − ux)2d
k≥1

hold.
Proof. We replace u r+1 by x in the proof of Proposition 5.6 and expand L by (34). 
The asymptotic behavior for the factorial moments of X n,r
„
and X n,r can now be extracted
quite straightforward by means of singularity analysis from the representation given in Propo-
sition 5.6.
Theorem 7. Let r ∈ N0 be fixed and consider n → ∞. Then the expected number of old leaves as
well as the expected number of nodes that are neither old leaves nor parents thereof in an r-fold
“old path”-reduced tree and the corresponding variances are given by the asymptotic expansions
1 (r + 3)(r + 1)
„
EX n,r = n+ + O(n−1 ), (47)
(r + 2) 2 6(r + 2)2
2(r + 1) (r 2 + 3r + 3)(r + 1)
EX n,r = n − + O(n−1 ),
(r + 2)2 3(r + 2)2
(r + 3)(r + 1)
„
VX n,r = n + O(1), (48)
3(r + 2)4
2(r 3 + 4r 2 + 6r + 6)(r + 1)
VX n,r = n + O(1).
3(r + 2)4
Additionally, for fixed d ≥ 2 the behavior of the factorial moments of X n,r
„
and X n,r is given by
d 1
„
EX n,r = + O(nd−1 ) (49)
(r + 2)2d

and
2d (r + 1)d d
d
EX n,r = n + O(nd−1 ), (50)
(r + 2)2d

respectively. All O-constants in this theorem depend implicitly on r.

5.3. Total number of old paths. Similarly to our approach for counting the total number of
paths required to construct a given tree from Section 3.2, we can also analyze the number of
“old path”-segments within a Prandom tree of size n. Formally, this corresponds to an analysis
of the random variable Sn := r≥0 X n,r„
.
36 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

Theorem 8. The expected number of “old path” segments within a uniformly random tree of
size n is given asymptotically by
€ π2 Š π2 1 π2
ESn = −1 n− − − + O(n−2 ) (51)
6 36 12 120n
for n → ∞.
Proof. As we have Sn = r≥0 X n,r
P „
, we can use (41) to write
X 1 1−u X u r+2
ESn = „
EX n,r = [z n ] .
r≥0
Cn−1 1 + u r≥0 (1 − u r+2 )2
The main part of this analysis consists of determining an appropriate expansion of the sum
in the last equation via the Mellin transform.
By setting u = e−t and by means of expanding via the geometric series, we find
X u r+2 X
λ(r+2)
X
= λu = λe−tλ(r+2) =: f (t).
r≥0
(1 − u r+2 )2
r,λ≥0 r,λ≥0

It is easy to determine the corresponding Mellin transform

f ∗ (s) = Γ (s)ζ(s − 1)(ζ(s) − 1)
with fundamental strip 〈2, ∞〉. The poles of f ∗ (s) are located at s ∈ {2, 1} ∪ −2N0 . As this
function behaves very nicely along vertical lines because of the exponential decay and the
polynomial growth of the gamma function and the zeta function, respectively, we can use the
inversion theorem to find Z 3+i∞
1
f (t) = f ∗ (s)t −s ds
2πi 3−i∞
for t → 0. Analyticity in a larger (complex) region can be obtained analogously to the
approach in the proof of Theorem 4.
Shifting the line of integration to Re(s) = −5 and collecting residues, we find
Z −5+i∞
X 1
f (t) = Res( f ∗ (s), s = p)t −p + f ∗ (s)t −s ds.
p∈{2,1,0,−2,−4}
2πi −5−i∞

As in the proof of Theorem 4, the integral can be estimated with an error of O(|t|5 ). However,
for the sake of simplicity, we will use the contribution from the singularity at s = −4 as the
expansion error. Effectively, we obtain
€ π2 Š 1 1 1 2
f (t) = − 1 t −2 − t −1 + − t + O(t 4 )
6 2 8 240
for t → 0. Multiplication with the factor 1−u
1+u , expansion of everything in terms of z → 1/4,
carrying out singularity analysis, and normalizing the result by dividing by Cn−1 yields the
result. 
 FRINGE ANALYSIS OF PLANE TREES RELATED TO CUTTING AND PRUNING 37

6. FUTURE WORK
It seems likely that similar results also hold for reductions where one can cut a different
structure as long as it is allowed to cut a single leaf. An example is cutting either single
leaves or cherries (a root with two children). At least a formulation as an operator as in (9)
seems possible in general. How much information about the moments and the central limit
theorem can be extracted from that may vary (as it varies in this article already). Also the
case of cutting old structures might be more difficult to handle in general.

REFERENCES
[1] David Callan, Kreweras’s Narayana number identity has a simple Dyck path interpretation, arXiv:1203.3999
[math.CO], 2012.
[2] William Y. C. Chen, Emeric Deutsch, and Sergi Elizalde, Old and young leaves on plane trees, European J.
Combin. 27 (2006), no. 3, 414–427.
[3] Nicolaas G. de Bruijn, Donald E. Knuth, and Stephen O. Rice, The average height of planted plane trees,
Graph theory and computing, Academic Press, New York, 1972, pp. 15–22.
[4] Mireille Vauchaussade de Chaumont, Nombre de Strahler des arbres, languages algébrique et dénombrement
de structures secondaires en biologie moléculaire, Doctoral thesis, Université de Bordeaux I, 1985.
[5] The SageMath Developers, SageMath Mathematics Software (Version 7.4), 2016, http://www.
sagemath.org.
[6] NIST Digital library of mathematical functions, http://dlmf.nist.gov/, Release 1.0.13 of 2016-09-16,
2016, Frank W. J. Olver, Adri B. Olde Daalhuis, Daniel W. Lozier, Barry I. Schneider, Ronald F. Boisvert,
Charles W. Clark, Bruce R. Miller and Bonita V. Saunders, eds.
[7] Michael Drmota, Random trees, SpringerWienNewYork, 2009.
[8] , Trees, Handbook of enumerative combinatorics, Discrete Math. Appl. (Boca Raton), CRC Press,
Boca Raton, FL, 2015, pp. 281–334.
[9] Philippe Flajolet and Andrew Odlyzko, The average height of binary trees and other simple trees, J. Comput.
System Sci. 25 (1982), no. 2, 171–213.
[10] , Singularity analysis of generating functions, SIAM J. Discrete Math. 3 (1990), 216–240.
[11] Philippe Flajolet, Jean-Claude Raoult, and Jean Vuillemin, The number of registers required for evaluating
arithmetic expressions, Theoret. Comput. Sci. 9 (1979), no. 1, 99–125.
[12] Philippe Flajolet and Robert Sedgewick, Analytic combinatorics, Cambridge University Press, Cambridge,
2009.
[13] Benjamin Hackl, Clemens Heuberger, and Helmut Prodinger, Reductions of binary trees and lattice paths
induced by the register function, arXiv:1612.07286 [math.CO], 2016.
[14] Benjamin Hackl, Sara Kropf, and Helmut Prodinger, Iterative cutting and pruning of planar trees, Proceed-
ings of the Fourteenth Workshop on Analytic Algorithmics and Combinatorics (ANALCO) (Philadelphia
PA), SIAM, 2017, pp. 66–72.
[15] Clemens Heuberger and Sara Kropf, Higher dimensional quasi-power theorem and Berry–Esseen inequality,
arXiv:1609.09599 [math.PR], 2016.
[16] Hsien-Kuei Hwang, On convergence rates in the central limit theorems for combinatorial structures, European
J. Combin. 19 (1998), 329–343.
[17] Svante Janson, Random cutting and records in deterministic and random trees, Random Structures Algo-
rithms 29 (2006), no. 2, 139–179.
[18] , Asymptotic normality of fringe subtrees and additive functionals in conditioned Galton-Watson trees,
Random Structures Algorithms 48 (2016), no. 1, 57–101.
[19] Rainer Kemp, A note on the stack size of regularly distributed binary trees, BIT 20 (1980), no. 2, 157–162.
38 B. HACKL, C. HEUBERGER, S. KROPF, AND H. PRODINGER

[20] Peter Kirschenhofer and Helmut Prodinger, Further results on digital search trees, Theoret. Comput. Sci. 58
(1988), no. 1-3, 143–154, Thirteenth International Colloquium on Automata, Languages and Program-
ming (Rennes, 1986).
[21] Amram Meir and John W. Moon, Cutting down random trees, J. Austral. Math. Soc. 11 (1970), 313–324.
[22] Alois Panholzer, Cutting down very simple trees, Quaest. Math. 29 (2006), no. 2, 211–227.
[23] Helmut Prodinger, The height of planted plane trees revisited, Ars Combin. 16 (1983), no. B, 51–55.
[24] Xavier Gérard Viennot, A Strahler bijection between Dyck paths and planar trees, Discrete Math. 246 (2002),
no. 1–3, 317–329, Formal Power Series and Algebraic Combinatorics 1999.
[25] Stephan Wagner, Central limit theorems for additive tree parameters with small toll functions, Combin.
Probab. Comput. 24 (2015), 329–353.
[26] Edmund T. Whittaker and George N. Watson, A course of modern analysis, Cambridge University Press,
Cambridge, 1996, Reprint of the fourth (1927) edition.
[27] Doron Zeilberger, A bijection from ordered trees to binary trees that sends the pruning order to the Strahler
number, Discrete Math. 82 (1990), no. 1, 89–92.

(Benjamin Hackl, Clemens Heuberger) INSTITUT FÜR MATHEMATIK, ALPEN-ADRIA-UNIVERSITÄT KLAGENFURT,

UNIVERSITÄTSSTRASSE 65–67, 9020 KLAGENFURT, AUSTRIA
E-mail address: benjamin.hackl@aau.at
E-mail address: clemens.heuberger@aau.at

(Sara Kropf) INSTITUTE OF STATISTICAL SCIENCE, ACADEMIA SINICA, 115 TAIPEI, TAIWAN
E-mail address: sarakropf@stat.sinica.edu.tw

(Helmut Prodinger) DEPARTMENT OF MATHEMATICAL SCIENCES, STELLENBOSCH UNIVERSITY, 7602 STELLEN-

BOSCH , SOUTH AFRICA
E-mail address: hproding@sun.ac.za